Income inequality and comorbid overweight/obesity and depression among a large sample of Canadian secondary school students: The mediator effect of social cohesion
School of Public Health, University of Alberta, 3-300 Edmonton Clinic Health Academy, 11405-87 Ave, Edmonton, AB T6G 1C9, Canada
Women and Children's Health Research Institute, University of Alberta, 5-083 Edmonton Clinic Health Academy, 11405 87 Avenue NW Edmonton, AB T6G 1C9, Canada
Brock University Department of Health Sciences, 1812 Sir Isaac Brock Way, St. Catharines, Ontario, L2S 3A1, Canada
School of Public Health Sciences, University of Waterloo, 200 University Avenue West Waterloo, Ontario N2L 3G1, Canada
School of Public Health, University of Alberta, Centre for Healthy Communities, Edmonton, AB T6G 2R3, Canada
Abstract
Background
Comorbid overweight/obesity (OWO) and depression is emerging as a public health problem among adolescents. Income inequality is a structural determinant of health that independently increases the risk for both OWO and depression among youth. However, no study has examined the association between income inequality and comorbid OWO and depression or tested potential mechanisms involved. We aimed to identify the association between income inequality and comorbid OWO and depression and to test whether social cohesion mediates this relationship.
Methods
We used data from the 2018–2019 Cannabis, Obesity, Mental health, Physical activity, Alcohol, Smoking and Sedentary behavior (COMPASS) project. Our sample was composed of 46,171 adolescents from 136 schools distributed in 43 census divisions in 4 provinces in Canada (Ontario, Alberta, British Columbia, and Quebec). Gender-stratified multilevel path analyses models were used to examine whether income inequality (Gini coefficient) was associated with comorbid OWO and depression and whether the association was mediated by school connectedness, a proxy measure for social cohesion.
Results
The direct effect between income inequality and OWO-depression comorbidity was not significant. However, income inequality was significantly associated with increased risk of comorbidity via social cohesion. One standard deviation increase in the Gini coefficient was associated with a 9% and 8% increase in the odds of comorbidity in females (OR=1.09; 95% CI=1.03, 1.16) and males (OR=1.08; 95% CI=1.03, 1.13).
Conclusion
Policies aimed at reducing income inequality, and interventions to improve social cohesion, may contribute to reducing the risk of OWO-depression comorbidity among adolescents.
Highlights
- •Comorbid overweight/obesity and depression is a major heath problem in adolescents.
- •We explored the mechanism by which income inequality increases comorbidity risk.
- •The prevalence of comorbidity averaged at 12.6% among the Canadian adolescents.
- •Income inequality increases the risk of comorbidity by eroding cohesion in schools.
- •The influence of income inequality is more pronounced in females relative to males.
Article notes
Untitled section
Received 2024 Feb 16; Revised 2024 Sep 7; Accepted 2024 Sep 8; Collection date 2024 Dec.
1.Introduction
Overweight/obesity (OWO) and depression in adolescents are major public health concerns. Recent school-based studies conducted from a convenience sample of Canadian adolescents (12–19 years) found that 40% of students report clinically-relevant symptoms of depression and/or anxiety (Williams et al., 2021) while 34.6% are at risk of overweight/obesity (Hunter et al., 2023). In addition to OWO and depression discretely being major health concerns, they frequently coexist among individuals, posing even greater threats to health (Fu et al., 2023). Comorbid OWO and depression is rapidly increasing in different populations with studies reporting of prevalence estimates ranging from 3.7% to 16% (Khanolkar & Patalay, 2021; Melton et al., 2021). Beyond the independent impact of depression and OWO on a person's health and wellbeing, individuals living with the comorbidity experience increased risk of cardiovascular diseases, functional disability, psychological distress, diabetes mellitus and higher health care utilization (Lin et al., 2022; Qin et al., 2023). Epidemiological and clinical studies have shown a plausible bidirectional causal relationship between OWO and depression due to shared psychosocial, metabolic and genetic mechanisms that contribute to the pathophysiology of the comorbidity (Milaneschi et al., 2019). This is corroborated by evidence that shows that 55% of individuals with obesity are at high risk of developing depression over time, while 58% of individuals with an early onset of depression are at high risk of developing obesity at some point in their lifetime (Luppino et al., 2010). Therefore, understanding the predictors of comorbidity may help in guiding public health interventions.
The epidemiology of the comorbid OWO and depression is an emerging area in research, with limited population-level studies available. Overall, comorbidity has been commonly reported among females (Scott et al., 2008), with some studies demonstrating heterogeneity of effects across females and males (Rajan & Menon, 2017). Comorbidity is also likely to develop in adolescence as this has been shown to be a critical period for the onset of OWO and mental health disorders (Bann et al., 2018; Patalay & Gage, 2019). Several other socio-demographic groups have also been reported to be associated with comorbidity, including individuals from low-income households, ethnic minority groups, frequent alcohol consumers, and individuals with low rates of physical activity (Chae et al., 2022; Haregu et al., 2020; Melton et al., 2021; Preiss et al., 2013). The influence of area-level factors on the occurrence of comorbidity has rarely been studied. Only two studies have investigated the role of socio-economic status and community size. Comorbidity was found to be common among socio-economically disadvantaged populations (Haregu et al., 2020) while there was no significant association reported in relation to community size (Chae et al., 2022). As such, there is an apparent gap in the current literature regarding structural determinants of comorbid OWO and depression.
One structural determinant of health that may be linked with comorbidity is income inequality, i.e., the disproportionate distribution of income in a specific group or area (Kawachi & Kennedy, 1999). Income inequality has been associated with adolescent BMI (Hunter et al., 2023; Lowe et al., 2023) and mental health (Benny et al., 2022; Pabayo et al., 2016) separately. Currently, there are no studies that have explored the relationship between income inequality and comorbid OWO and depression. However, comorbidity literature suggests that diseases could co-occur due to ‘a shared pattern of influence’ (Fu et al., 2023; Valderas et al., 2009). Therefore, it can be postulated that income inequality, by virtue of being a common risk factor for both depressive symptoms and OWO (Lowe et al., 2023; Pabayo et al., 2016), could also be associated with higher risk of comorbidity.
1.1.Theoretical background
Drawing from the social determinants of health framework, income inequality can be conceptualized as a structural determinant that can directly or indirectly influence adolescent health in various ways (Goodman, 2008; Herge et al., 2013; Viner et al., 2012). The seminal works of Kawachi and colleagues (Kawachi & Kennedy, 1999; Wilkinson, 1999) identify three mechanisms through which income inequality may potentially influence health outcomes. First is the social comparison pathway that suggests that negative health outcomes are results of invidious comparison of social status (Kawachi & Kennedy, 1999). This mechanism posits that societies have status hierarchies in which individual's compete for status to increase their social standing within the society (Prag et al., 2013; Wilkinson, 1999). Failure to achieve the societal ‘ideal status’ invites negative appraisals to one's status which leads to chronic stress that negatively affects ones health (Dickerson & Kemeny, 2004; Prag et al., 2013). Second is the neo-materialist pathway that proposes that income distribution shapes investments in human capital, health and social infrastructure (Lynch et al., 2004). Societies with high inequalities are thus characterised with underinvestment in human capital and other public infrastructure due to limited collective resources to facilitate such investments (Lynch et al., 2004; Prag et al., 2013). This in turn will shape the quality and access of health-promoting public infrastructure such as schools, health care services and occupational opportunities that can help to boost individuals social mobility (Lynch et al., 2004).
The third pathway that is complementary to the neo-materialist mechanism is the social cohesion and social capital pathways. Literature suggests that income inequality erodes social cohesion and social capital (Kawachi & Kennedy, 1999). Social cohesion is key in building trust and cooperation that in turn buffer the community from behaviors that may be deleterious to their health (ibid.). Therefore, declines in social cohesion could lead to the feeling of social isolation and individual's disconnection from the community (Santini et al., 2020). Potential consequences of suppressed communal engagements may be an increase in depressive symptoms (Pabayo et al., 2016), decreases in community physical activity and inadequate access to healthy diets (Clément et al., 2021) which may synergistically lead to comorbidity. Social cohesion within the school environment could be playing an important role in shaping adolescent health since the majority of adolescents spend a substantial amount of their time in schools. Cohesion in terms of a bonding and feeling a sense of belonging in school has been shown to be a protective factor against negative health outcomes and delinquent behaviors (Nasir et al., 2011; Quader et al., 2022; Zhang et al., 2023). A proxy measure that has been used in previous research among school-going adolescents to capture social cohesion is school connectedness (Benny et al., 2022). Certain aspects of the commonly used school connectedness scale such as fair treatment, closeness and feelings of belonging (McNeely et al., 2002; Patte et al., 2021) resonate well with the concepts of mutual trust and a sense of belonging engendered in the social cohesion construct (Kawachi & Subramanian, 2014; Pei et al., 2020), hence making school connectedness a valid proxy for social cohesion.
The effect of income inequality in shaping health outcomes is by no means similar across all population subgroups. Females, relative to males, have been found to show higher sensitivity to income inequality in various health outcomes including weight gain, sedentary lifestyle, smoking and depression (Diez-Roux et al., 2000; Pabayo et al., 2016; Patel et al., 2018). Cultural gender norms, including sociocultural appearance ideals, may offer explanations as to why such disproportionate effects are likely to occur. Income inequality may heighten attention to other social hierarchies and comparisons, including by weight and gender, and the perceived need to conform to gendered norms. For example, females residing in areas with high income inequality are more likely to be stigmatized and discriminated based on their weight when compared to males (Li, 2024), and weight stigma is known to predict both weight gain and depression (Greenleaf et al., 2017). Additionally, female adolescents are also known to internalize their problems as a coping mechanism for stressful environments (Hankin et al., 2007; Pabayo et al., 2016). Given that internalizing problems have been shown to elevate the risk of both OWO (Preiss et al., 2013) and depression (Zahn-Waxler et al., 2000), there is a high probability that comorbidity will disproportionately affect female adolescents relative to their male counterparts.
To our knowledge, no prior studies have investigated the association between income inequality and comorbidity. Therefore, this study aims to test this relationship empirically; the objectives are: 1) To identify the association between income inequality and OWO and depression comorbidity; and 2) to investigate whether social cohesion (school connectedness) mediates the relationship between income inequality and comorbidity. We hypothesized that census division income inequality will have a direct effect with comorbidity. Our direct effect hypothesis relies on two mechanisms. i.e., the invidious social comparison and neo-materialist mechanisms. Given that in this study we were unable to test the potential mediation of these two mechanisms, we anticipated that these effects would be reflected on the direct association between income inequality and comorbidity. We further hypothesized that social cohesion, would significantly mediate the relationship between census division income inequality and comorbidity. Our indirect effect hypothesis is pegged on the mechanism underlying the impact of income inequality in eroding social cohesion (Benny et al., 2022; Kawachi & Kennedy, 1999). Finally, we hypothesized that the direct and/or mediated relationship between income inequality and comorbidity would be stronger in females than males as females internalizing behaviours in response to stressful environments (Pabayo et al., 2016) may elevate their risk of comorbidity.
2.Methods
This study used secondary data from the Cannabis, Obesity, Mental health, Physical activity, Alcohol, Smoking and Sedentary behavior (COMPASS) project. Established in 2012, the COMPASS project is an ongoing prospective cohort study that collects survey data annually from secondary schools and students attending those schools (grades 9 to 12) (Leatherdale et al., 2014). COMPASS covers four Canadian provinces: Ontario, Alberta, British Columbia, and Quebec. COMPASS utilizes a convenience sampling approach targeting high schools that permit use of active information–passive consent protocols (Hompson-Haile et al., 2013). All study protocols for the COMPASS project were approved by the University of Waterloo Office of Research Ethics (#30118), University of Alberta Ethics (#RE0050375) and the appropriate school board committees of participating schools (Leatherdale et al., 2014). Further details on the COMPASS study methods and protocols are available online (www.compass.uwaterloo.ca).
The current study utilized cross-sectional data from the 2018–2019 school year (wave 7). In wave 7 of the COMPASS project, full school samples of students completed a paper-and-pencil questionnaire during class time. The analysis was limited to participants who had complete data on all variables of interest. Subsequently, the case complete COMPASS data was linked to an area-level (Census divisions) dataset obtained from the 2016 Canadian Census that contained measures of income inequality and other covariates of interests. Census divisions are geographically demarcated areas situated between municipalities and provinces in Canada (Statistics Canada, 2018). They are established to facilitate regional planning and are used by government to conduct the periodic census (Statistics Canada, 2018). Out of 74,501 students that participated in the 2018–2019 survey, 46,733 (62.7%) from 136 schools nested within 43 census divisions in 4 provinces in Canada (Ontario, Alberta, British Columbia, and Quebec) had complete data and thus formed our analytic sample for this study.
2.1.Measures
2.1.1.Outcome
Risk of OWO was assessed using body mass index (BMI) calculated as weight in kilograms divided by height in squared meters. Before calculating BMI values, all reported measures for height and weight were multiplied by sex-specific correction factors for females (height=1.00116 and weight=1.04779) and males (height=1.00058 and weight=1.02998) to correct for biases associated with self-reported anthropometric values (Leatherdale & Laxer, 2013). BMI values were transformed into age- and sex-standardized z-scores (WHO, 2007b) and classified based on the World Health Organization's (WHO) guidelines as follows: underweight (<-2), normal weight (−2 to ≤1), overweight (>1 to ≤2) and obesity (>2) (WHO, 2007a). Participants with implausible z-BMI scores of <-5 or >5 were excluded from the study. Additionally, we also excluded underweight participants (<2%) since they did not form an adequate sample size for estimating statistical differences.
Participants' depressive symptoms were measured using a 10-item Center for Epidemiologic Studies Depression Scale Revised (CESD -R). The CESD-R scale involves assessing the frequency of occurrence of 10 depressive symptoms over the previous week (Cronbach's α=0.82) (Van Dam & Earleywine, 2011). Plausible scores range from 0 to 30 with higher scores indicative of more frequent depressive symptoms. A binary variable was constructed to assess whether participants had clinically-relevant depression symptoms (CESD-R score ≥10) or not (CES-D score <10) (Van Dam & Earleywine, 2011).
The outcome of interest, comorbidity, was created by combining the BMI and depression variables as follows: i) Neither classified as having OWO nor depression (reference group); ii) classified as having OWO only; iii) depression only; and iv) comorbid (i.e., classified as having both OWO and depression).
2.1.2.Exposure
Income inequality was assessed using the Gini coefficient. The Gini coefficient was computed at the Census Division (CD) level using household income data from the 2016 Canadian Census. A Lorenz curve was created by plotting the cumulative proportion of the household income against the cumulative proportion of the population (Benny et al., 2023). The Gini coefficient was calculated by dividing the area between the line of equality and the Lorenz curve (A) by the total area beneath the perfect line of equality (A + B) (Fig. 1). A detailed computation of the Gini coefficient can be found elsewhere (Dorfman, 1979). Values of the Gini coefficient ranged from 0 (equal distribution) to 1 (unequal distribution). The Gini coefficient variable was z-transformed for ease of interpretation of the results.
2.1.3.Mediating variable
This study used the school connectedness scale as a proxy measure of the social cohesion construct. Derived from previous studies on school going adolescents (McNeely et al., 2002), COMPASS surveys use a 6-item scale to assess school connectedness (Cronbach's α=0.82). Students were asked to provide responses on a 4-point Likert scale to the following six statements: “I feel close to people at my school”, “I feel I am part of my school”, “I am happy to be at my school”, “I feel the teachers at my school treat me fairly”, “I feel safe in my school”, and “Getting good grades is important to me”. Responses were summed (scores between 6 and 24) and higher scores were indicative of greater connectedness. In line with previous research (Benny et al., 2022), we excluded the item “Getting good grades is important to me” from the connectedness scale as it is not reflective of the social cohesion construct espoused in this study. Therefore, our new measure consisted of five items with score between 4 and 20.
2.1.4.Covariates
In COMPASS surveys, respondents were asked to report whether they were female or male. Literature suggests that the ‘female’ or ‘male’ responses in surveys may refer to biological (sex) and social (gender) identity (Johnson et al., 2009). In the present study, we used gender (henceforth) since the study focuses on social processes that may lead to health inequities. Covariates at the individual level included age (12–19 years), ethnicity (white or people of colour), the amount of money that the student has available to spend on themselves on a weekly basis ($0, $1-$5, $6-$10, $11-$20, $21-$40, $41-$100, >$100, and don't know), alcohol consumption (none, <once per month, monthly, weekly, not stated), and amount of time spent in vigorous physical activity per day (<60 min per day, ≥60 min per day, not stated). At the CD-level, the covariates included provinces (Alberta, British Columbia, Ontario, and Quebec), median after-tax household income and population size.
2.1.5.Statistical analysis
Initial analysis involved estimating the prevalence of OWO and depression. Descriptive statistics were calculated by comparing the prevalence of the outcome variable across the exposure, mediator, and the covariates. Our main analysis involved the use of a multilevel (two-level) path analysis approach to examine the hypothesized effect of income inequality on comorbidity. Multilevel models were selected due to the hierarchical nature of the dataset where students and schools were nested within CDs, whereas the path analysis framework was suitable because it allows for the examination of both the direct and indirect (mediated) associations of the exposure.
Data cleaning and coding of variables were done using STATA v16.0, while path analysis was performed using Mplus v8.8. With the study outcome being a nominal variable, multinomial regression would have been the ideal model for this analysis. However, multinomial regression models within the framework of path analysis have not been validated. Therefore, we used a two-level logistic regression model, where level-1 was composed of individual level variables and level-2 being measurements at the CD. For computational efficiency with Mplus (Muthén et al., 2011), our outcome variable was split into a set of three binary outcomes as follows: 1) Neither OWO or depression vs. OWO-depression comorbidity; 2) Neither OWO or depression vs. OWO only; and 3) Neither OWO or depression vs. depression only. Furthermore, the analysis was stratified by gender to examine whether the effect of income inequality was heterogeneous between females and males. All models were fitted using the maximum likelihood robust (MLR) estimator, which is robust to non-normality. For ease of interpretation, the regression coefficient estimates were exponentiated and interpreted as odds ratio (OR) at 5% significance level.
2.1.5.1.Assessing model fit
Model fitness for path analysis models is commonly assessed by a combination of relative indices such as Tucker-Lewis index (TLI) and absolute fit indices such as Standardized Root Mean Square Residuals (SRMR). However, when using multilevel models, these fit indices are not available. In such cases, the use of likelihood ratio tests (nested models) and information criteria (non-nested models) have been recommended (Curran et al., 2010). Therefore, we assessed model fitness using the log-likelihood tests. This was performed by comparing the unadjusted mediated models (model 1) to the fully adjusted mediated model (model 2), with the models with a higher value indicating adequate fit.
3.Results
3.1.Descriptive statistics
The prevalences of overweight and obesity were 21.6% and 8.0% among females and 23.7% and 12.6% among males, respectively. The prevalence of underweight was <2% in each gender group. The estimated prevalence of depression was 46.2% among females and 25.5% among males. After excluding the underweight participants (n=562), the final analytic sample was composed of 46,171 adolescents from 136 schools distributed in 43 census divisions (Table 1). Nearly half of the adolescents had neither OWO or depression (43.4%), 20.7% had OWO alone, 23.3% had depression alone and 12.6% had comorbid OWO-depression. Among those with comorbidity, the majority were females (60.9%), white students (66.5%), those with available weekly spending money of >$100 (24.3%), alcohol consumers (64.4%), students who undertook <60 min of vigorous physical activity per day (60.1%) and residents of Ontario (49.6%).
| Individual Characteristics | Neither OWO- Depression | OWO Only | Depression Only | Comorbid OWO and Depression |
|---|---|---|---|---|
| Frequency n=46,171 | 20,043 | 9541 | 10,735 | 5852 |
| Age, mean (std) | 15.19 (±1.49) | 15.12 (±1.47) | 15.51 (±1.36) | 15.49 (±1.39) |
| Gender, % | ||||
| Male | 10,911 (54.44) | 6112 (64.06) | 3508 (32.68) | 2291 (39.15) |
| Female | 9132 (45.56) | 3429 (34.94) | 7227 (67.32) | 3561 (60.85) |
| Ethnicity, % | ||||
| White | 15,072 (75.20) | 7053 (73.92) | 7311 (68.10) | 3891 (66.49) |
| People of colour | 4971 (24.80) | 4971 (26.08) | 3424 (31.90) | 1961 (33.51) |
| Weekly Spending Money, % | ||||
| $0 | 2711 (13.53) | 1374 (14.40) | 1498 (13.95) | 895 (15.29) |
| $1-$5 | 1054 (5.26) | 460 (4.82) | 587 (5.47) | 298 (5.09) |
| $6-$10 | 1320 (6.59) | 585 (6.13) | 682 (6.35) | 334 (5.71) |
| $11-$20 | 2347 (11.71) | 1036 (10.86) | 1282 (11.94) | 656 (11.21) |
| $21-$40 | 2271 (11.33) | 1114 (11.68) | 1245 (11.60) | 650 (11.11) |
| $41-$100 | 2595 (12.95) | 1221 (12.80) | 1535 (14.30) | 804 (13.74) |
| >$100 | 4105 (20.48) | 2127 (22.29) | 2286 (21.29) | 1424 (24.33) |
| Don't know | 3640 (18.16) | 1624 (17.02) | 1620 (15.09) | 791 (13.52) |
| Alcohol Consumption, % | ||||
| None | 9527 (47.53) | 4397 (46.09) | 3902 (36.35) | 2060 (35.20) |
| <Once per month | 4323 (21.57) | 2056 (21.55) | 2602 (24.24) | 1419 (24.25) |
| Monthly | 4512 (22.51) | 2177 (22.82) | 3086 (28.75) | 1688 (28.84) |
| Weekly | 1592 (7.94) | 870 (9.12) | 1104 (10.28) | 661 (11.30) |
| Not stated | 89 (0.5) | 41 (0.43) | 41 (0.38) | 24 (0.41) |
| Physical Activity, % | ||||
| <60 min per day | 11,923 (59.49) | 5413 (56.73) | 6812 (63.46) | 3517 (60.10) |
| ≥60 min per day | 7912 (39.48) | 4047 (42.42) | 3823 (35.61) | 2272 (38.82) |
| Not stated | 208 (1.04) | 81 (0.85) | 100 (0.93) | 63 (1.08) |
| Social Cohesion | 16.03 (±2.59) | 15.93 (±2.71) | 13.77 (±2.87) | 13.51 (±2.95) |
| Income Inequality (Gini) | 0.37 (±0.03) | 0.37 (±0.03) | 0.37 (±0.03) | 0.37 (±0.03) |
| Provinces | ||||
| Alberta | 860 (4.29) | 463 (4.85) | 512 (4.77) | 347 (5.93) |
| British Columbia | 2660 (13.27) | 968 (10.15) | 1793 (16.70) | 740 (12.65) |
| Ontario | 7466 (37.25) | 4077 (42.73) | 4666 (43.47) | 2901 (49.57) |
| Quebec | 9057 (45.19) | 4033 (42.27) | 3764 (35.06) | 1864 (31.85) |
| Median after-tax income | 59006.14 (±8901.52) | 58850.40 (±8714.41) | 60204.00 (±8728.77) | 60207.80 (±8653.38) |
| Population Size | 546565.60 (±667367.30) | 464146.80 (±576972.30) | 613023.60 (±724,566.00) | 514797.80 (±642240.90) |
3.2.Path analysis
3.2.1.Comorbidity
In relation to model fitness, the log-likelihood values for covariate adjusted gender-stratified models (female=−36641.964, males=−37134.965) were higher than those of crude mediated models (female=−36752.922, males=−37218.505) indicating an adequate fit for our final models. Fig. 2 presents the gender stratified unstandardized coefficients of the direct and indirect paths in the multilevel path analysis models. For either gender, the direct effect was not statistically significant. However, the indirect (mediated) effects were statistically significant. A unit increase in social cohesion score was associated with lower risk of comorbid OWO-depression in females (β=−0.34; 95% CI=−0.36, −0.32) and males (β=−0.29; 95% CI=−0.31, −0.27) (Table S1, Supplementary document). However, this protective effect was nullified by the effect of income inequality on social cohesion where a one-standard deviation (SD) increases in Gini coefficient resulted to a decrease in social cohesion in females (β=−0.25; 95% CI=−0.43, −0.08) and males (β=−0.25; 95% CI=−0.41, −0.10) respectively. In sum, relative to students with neither condition, one-SD increase in Gini coefficient via the social cohesion mediated pathway was associated with a 9% and 8% increase in the odds of comorbid OWO-depression in females (OR=1.09; 95% CI=1.03, 1.16) and males (OR=1.08; 95% CI=1.03, 1.13) respectively (Table 2).
| Variable | Female OR (95% CI) | Male OR (95% CI) | |
|---|---|---|---|
| Direct Effect | Income Inequality | 0.96 (0.89 1.03) | 1.05 (0.92 1.19) |
| Indirect Effect | Income Inequality → Social Cohesion | 1.09 (1.03 1.16) | 1.08 (1.03 1.13) |
| Covariates | Individual level | ||
| Age | 0.94 (0.91 0.98) | 1.02 (0.97 1.09) | |
| Ethnicity | |||
| White | 1.00 | 1.00 | |
| People of colour | 1.53 (1.23 1.90) | 1.51 (1.32 1.72) | |
| Weekly spending money | |||
| $0 | 1.00 | 1.00 | |
| $1 to $5 | 0.98 (0.80 1.19) | 0.89 (0.74 1.06) | |
| $6 to $10 | 0.90 (0.76 1.08) | 0.69 (0.57 0.84) | |
| $11 to $20 | 0.85 (0.71 1.02) | 0.91 (0.76 1.09) | |
| $21 to $40 | 0.78 (0.64 0.95) | 0.91 (0.79 1.05) | |
| $41 to $100 | 0.81 (0.69 0.94) | 0.79 (0.67 0.94) | |
| >$100 | 0.90 (0.78 1.04) | 0.83 (0.71 0.97) | |
| Don't know | 0.70 (0.59 0.83) | 0.73 (0.64 0.85) | |
| Alcohol consumption | |||
| None | 1.00 | 1.00 | |
| <Once per month | 1.53 (1.38 1.68) | 1.26 (1.12 1.42) | |
| Monthly | 1.89 (1.60 2.24) | 1.62 (1.36 1.92) | |
| Weekly | 1.92 (1.52 2.44) | 1.75 (1.39 2.20) | |
| Not stated | 0.87 (0.38 1.98) | 1.50 (0.74 3.02) | |
| Physical activity | |||
| <60 min per day | 1.00 | 1.00 | |
| >60 min per day | 1.11 (1.02 1.21) | 0.89 (0.79 1.00) | |
| Not stated | 0.88 (0.61 1.28) | 0.91 (0.56 1.49) | |
| CD Level | |||
| Provinces | |||
| Ontario | 1.00 | 1.00 | |
| Alberta | 1.05 (0.79 1.39) | 0.90 (0.60 1.35) | |
| British Columbia | 0.87 (0.73 1.04) | 0.89 (0.69 1.16) | |
| Quebec | 0.64 (0.57 0.72) | 0.55 (0.43 0.70) | |
| Median after tax income | 0.95 (0.93 0.97) | 1.00 (0.92 1.09) | |
| Population size | 0.98 (0.96 1.00) | 0.99 (0.97 1.00) | |
3.2.2.Comparison between overweight/obesity alone, depression alone and comorbidity
Path analyses findings for the depression alone and OWO alone groups are presented in the supplemental document (Tables S2-S5 and Figs. S1 and S2). The effects observed among the depression alone group were like that of the comorbid OWO and depression, albeit with smaller effect sizes. One-SD increase in Gini was indirectly (via school connectedness) associated with higher odds of depression in females (OR=1.08; 95% CI=1.03, 1.14) and males (OR=1.07; 95% CI=1.02, 1.11), compared to a group of students with neither condition (Table S2). For the OWO alone group, the indirect effect of income inequality was only observed among females where a one-SD increase in Gini resulted to a modest increase in the odds of OWO (OR=1.01; 95% CI=1.00, 1.02) (Table S4).
4.Discussion
In this multilevel path analysis study, we found that there was no direct association between income inequality and comorbidity. However, income inequality was indirectly associated with comorbid OWO and depression via the social cohesion mediated pathway. In our stratified analyses, we observed that the regression coefficients of this association was slightly stronger among females than in males. Finally, we also found that the observed significant indirect effect of income inequality was comparable to that of individuals with depression only, although the regression coefficients for comorbidity was slightly stronger.
There are various mechanisms through which income inequality may contribute to adverse health outcomes. Findings from this study suggests that erosion of social cohesion (Kawachi & Kennedy, 1999) (school connectedness) could be one of the pathway through which income inequality increases the risk of OWO-depression comorbidity among school going adolescents. School connectedness has been shown to be an important factor that influences adolescents’ health and well-being, including conferring protection against depression and excess weight gain (Quader et al., 2022; Zhang et al., 2023). However, residential areas with unequal distribution of income are unlikely to provide conditions that promote connectedness. Communities affected with large income inequalities are likely to experience disruptions in social and political systems, which in turn may lead to increases in crime, anxiety and stress (Pop et al., 2013). Furthermore, highly unequal areas are characterized with underinvestment in social infrastructure and human capital, limiting access to resources that are crucial in promoting human health (Kawachi & Kennedy, 1999; Kawachi & Subramanian, 2014). Such constraints could be reflected in inadequate funding of school boards and thus school programs that promote connectedness, such as extracurricular activities, safety programs, counselling services, and training of educators on emerging challenges such as mental health (Nasir et al., 2011), are likely to be scarce in unequal areas. Therefore, the risk of comorbidity is likely to be high in such divested environments.
The effects of income inequality seemed to be slightly pronounced in female than male students. This finding is in line with previous studies that have shown higher risk of comorbidity among females (Chae et al., 2022; Lin et al., 2022), which is likely driven by the high prevalence of depression in females relative to males. This gender difference may be due to differences in coping mechanisms that may expose females more to high-risk behaviors. Girls dealing with stress are likely to withdraw socially and are more likely to seek solace in unhealthy eating, accompanied with a physically inactive lifestyle (Mooreville et al., 2014; Quader et al., 2022). Indulgence in such risky behaviors is likely to increase the risk of comorbidity.
Additionally, this study also found that the regression coefficients of income inequality on comorbidity was slightly stronger compared to that of income inequality and depression alone across the gender groups. This is not surprising as previous studies have also reported an exaggerated effect in relation to the risk factors of comorbid OWO and depression. For instance, Khanolkar and Patalay (2021) using pooled data from two British national birth cohorts found that adolescents (16 years) classified as belonging to a disadvantaged childhood social class (i.e., father's social class being partly skilled and unskilled) were associated with a much higher risk of comorbid OWO and mental-ill health (RRR=2.04; 95% CI=1.54, 2.72) over and above the risk of OWO (RRR=1.39; 95% CI=1.21, 1.58) and mental-ill health (RRR=1.36; 95% CI=1.22, 1.51) separately. From our analysis, it is clear that this elevated risk of comorbidity is highly driven by the erosion of school connectedness. This suggests that deterioration of school connectedness due to high income inequality could be impacting a common pathway for both OWO and depression that cumulatively leads to high risk of comorbidity. However, with lack of literature to back our hypothesis, we recommend future studies to investigate the sequential precedence of mechanisms that underlies the relationship between school connectedness and comorbidity.
Contrary to our hypothesis, this study did not find any significant direct association between income inequality and comorbidity. Our anticipation was that a residual effect from the two hypothesized mechanisms (social comparison and neo-materialist pathways) that were untested in this study would have been reflected in this direct path. The reasons as to why we observed this null finding are highly speculative. First, the study findings may be suggesting that in this specific population of Canadian adolescents, the relationship between income inequality and comorbid OWO and depression is potentially not mediated by social comparison and/or neo-materialism mechanisms. Second, the mechanisms underlying the influence of income inequality on health outcome are not mutually exclusive. For example, there is a reciprocal relationship between neo-materialism and social cohesion in that in schools that have inadequate investments in co-curricular activities, human capital and other social programs may as well contain students who feel less connected and accepted within their school environments (Benny, 2023; Jenson & Saint-Martin, 2003; Kawachi & Subramanian, 2014). Therefore, there is a likelihood that the hypothesized effect of the neo-materialism was usurped by the school connectedness pathway. However, these explanations are not certain given the data limitations we have with the current study. Future studies using longitudinal data may be able to offer better clarity of the influence of the untested mechanisms.
The emergence of comorbidity among adolescents indicates that siloed approach of public health interventions aimed at reducing the risk of OWO and depression separately may be ineffective in reducing health morbidities. Integrated approaches that target root causes of comorbid conditions are likely to be more effective (Khanolkar & Patalay, 2021). Findings from this study indicate that policies that focus on the reduction of census-division income inequality may have a long-lasting impact in addressing comorbidity since such ‘upstream’ factors also influence the operation of ‘downstream’ risk factors. Development of economic policies that aim to achieve fair distribution of income such as progressive taxation and social policies geared towards reducing income inequities such as equitable educational opportunities can contribute substantially in reducing the risk of comorbidity (Patel et al., 2018). Additionally, our results also reveal that certain aspects of the school environment can be leveraged with the aim of promoting adolescent health. Measures leverage on the power of school connectedness in creating social trust and reducing social isolation in areas characterized with unequal distribution of income can contribute to the reduction of the risk of comorbidity.
Several limitations should be considered when interpreting the findings from this study. First, even though we had a large sample size, nearly one third of survey participants did not report their weight, height and depressive symptoms values (Doggett et al., 2022) hence the exclusion of these participants in our analysis may have introduced some bias in our findings. To assess the extent of this potential bias, we conducted two exploratory tests: 1) we compared the proportions of the non-response and complete case samples across the levels of each covariates using chi-square and 2) we conducted a logistic regression predicting the odds of non-response versus case complete, using the study covariates as the predictors. Our findings suggested that non-respondents were more likely to be younger (<15 years), identified as an ethnic minority, had low weekly spending money (5 dollars or less) and those who had never taken alcohol. However, proportions of non response sample compared to the case complete sample was not substantially different across the levels of the covariates, hence we believe that the findings of our study are still useful despite these biases. Second, since COMPASS surveys use convenience sampling, our findings cannot be fully generalized to the entire Canadian population. Third, the study used cross sectional data and therefore causal inferences cannot be made based on our findings. Finaly, our findings may also be biased due to residual confounding since we were unable to fully adjust for known confounders such as house income.
In conclusion, this study found that income inequality via social cohesion (school connectedness) is associated with higher risk of comorbid OWO and depression among school going adolescents, with females being slightly more affected than males. Moreover, adolescents attending schools in areas with high income inequality are potentially at higher risk of developing comorbidity than OWO or depression alone. Future studies can build from this study by investigating this relationship using longitudinal data to determine the impact of areas income inequality on comorbidity from a life course dimension.
Ethical statement
This study used data from the Cannabis, Obesity, Mental health, Physical activity, Alcohol, Smoking and Sedentary behavior (COMPASS) project. The current study was approved by the University of Waterloo Office of Research Ethics (#30118), the University of Alberta Ethics (#RE0050375) as well as the school board committees of participating schools. However, the decision to submit the manuscript and the views reported in this paper are those of the authors and does not represent the official views of the University of waterloo or University of Alberta.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements of support and assistance
The COMPASS study has been supported by a bridge grant from the CIHR Institute of Nutrition, Metabolism and Diabetes (INMD) through the “Obesity – Interventions to Prevent or Treat” priority funding awards (OOP-110788; awarded to SL), an operating grant from the CIHR Institute of Population and Public Health (IPPH) (MOP-114875; awarded to SL), a CIHR project grant (PJT-148562; awarded to SL), a CIHR bridge grant (PJT-149092; awarded to KP/SL), a CIHR project grant (PJT-159693; awarded to KP), and by a research funding arrangement with Health Canada (#1617-HQ-000012; contract awarded to SL), a project grant from the CIHR Institute of Population and Public Health (IPPH) (PJT-180262; awarded to SL and KP). The COMPASS-Quebec project additionally benefits from funding from the Ministère de la Santé et des Services sociaux of the province of Québec, and the Direction régionale de santé publique du CIUSSS de la Capitale-Nationale. For this analysis, funding was received from CIHR Operating Grant: Data Analysis Using Existing Canadian Databases and Cohorts, #489013. KAP is the Canada Research Chair in Child Health Equity and Inclusion. RP is a Tier II Canada Research Chair in Social and Health Equities. SH is supported by a Women and Children’s Health Research Institute Postdoctoral Fellowship Award. This Postdoctoral Fellowship has been funded by the Alberta Women’s Health Foundation and the Stollery Children’s Hospital Foundation through the Women and Children’s Health Research Institute.
We would like to express our gratitude to the COMPASS study for providing access to the data that was used to conduct this study. Further, we would like to acknowledge the University of Alberta for providing the library resources needed for the completion of this study.
Footnotes
Footnote Group
Appendix A.Supplementary data
The following is the Supplementary data to this article:
Data availability
COMPASS study data is available upon request through completion and approval of an online form: https://uwaterloo.ca/compass-system/information-researchers/data-usage-application The datasets used during the current study are available from the corresponding author on reasonable request.
References
Untitled section
References
- Bann D., Johnson W., Li L., Kuh D., Hardy R. Socioeconomic inequalities in childhood and adolescent body-mass index, weight, and height from 1953 to 2015: An analysis of four longitudinal, observational, British birth cohort studies. The Lancet Public Health. 2018;3(4):e194–e203. doi: 10.1016/S2468-2667(18)30045-8.
- Benny C. Investigating the association between income inequality and mental health and deaths of despair in Canadian youth [University of Alberta] 2023.
- Benny C., Patte K.A., Veugelers P., Leatherdale S.T., Pabayo R. Income inequality and depression among Canadian secondary students: Are psychosocial well-being and social cohesion mediating factors? SSM - Population Health. 2022;17 doi: 10.1016/j.ssmph.2021.100994.
- Benny C., Patte K.A., Veugelers P.J., Senthilselvan A., Leatherdale S.T., Pabayo R. A longitudinal study of income inequality and mental health among Canadian secondary school students: Results from the Cannabis, obesity, mental health, physical activity, alcohol, smoking, and sedentary behavior study (2016–2019) Journal of Adolescent Health. 2023;73(1):70–78. doi: 10.1016/j.jadohealth.2023.01.028.
- Chae W.R., Schienkiewitz A., Du Y., Hapke U., Otte C., Michalski N. Comorbid depression and obesity among adults in Germany: Effects of age, sex, and socioeconomic status. Journal of Affective Disorders. 2022;299:383–392. doi: 10.1016/j.jad.2021.12.025.
- Clément M., Levasseur P., Seetahul S., Piaser L. Does inequality have a silver lining? Municipal income inequality and obesity in Mexico. Social Science & Medicine. 2021;272 doi: 10.1016/j.socscimed.2021.113710.
- Curran P.J., Obeidat K., Losardo D. Twelve frequently asked questions about growth curve modeling. Journal of Cognition and Development. 2010;11(2):121–136. doi: 10.1080/15248371003699969.
- Dickerson S.S., Kemeny M.E. Acute stressors and cortisol responses: A theoretical integration and synthesis of laboratory research. Psychological Bulletin. 2004;130(3):355–391. doi: 10.1037/0033-2909.130.3.355.
- Diez-Roux A.V., Link B.G., Northridge M.E. A multilevel analysis of income inequality and cardiovascular disease risk factors. Social Science & Medicine. 2000;50(5):673–687. doi: 10.1016/s0277-9536(99)00320-2.
- Doggett A., Chaurasia A., Chaput J.-P., Leatherdale S.T. Learning from missing data: Examining nonreporting patterns of height, weight, and BMI among Canadian youth. International Journal of Obesity. 2022;46(9):1598–1607. doi: 10.1038/s41366-022-01154-8.
- Dorfman R. A formula for the Gini coefficient. The Review of Economics and Statistics. 1979;61(1):146–149. doi: 10.2307/1924845.
- Fu X., Wang Y., Zhao F., Cui R., Xie W., Liu Q., Yang W. Shared biological mechanisms of depression and obesity: Focus on adipokines and lipokines. Aging. 2023;15(12):5917–5950. doi: 10.18632/aging.204847.
- Goodman E. In: Children and adolescents BT - handbook of childhood and adolescent obesity. Jelalian E., Steele R.G., editors. Springer; US: 2008. Socioeconomic factors related to obesity; pp. 127–143.
- Greenleaf C., Petrie T.A., Martin S.B. Exploring weight-related teasing and depression among overweight and obese adolescents. European Review of Applied Psychology. 2017;67(3):147–153. doi: 10.1016/j.erap.2017.01.004.
- Hankin B.L., Mermelstein R., Roesch L. Sex differences in adolescent depression: Stress exposure and reactivity models. Child Development. 2007;78(1):279–295. doi: 10.1111/j.1467-8624.2007.00997.x.
- Haregu T.N., Lee J.T., Oldenburg B., Armstrong G. Comorbid depression and obesity: Correlates and synergistic association with noncommunicable diseases among Australian men. Preventing Chronic Disease. 2020;17:E51. doi: 10.5888/pcd17.190420.
- Herge W., Landoll R., La Greca A. Encyclopedia of behavioral medicine. 2013. Center for epidemiologic studies depression scale (CES-D) pp. 366–367.
- Hompson-Haile A., Bredin C., Leatherdale S. Rationale for using an active-information passive-consent permission protocol in COMPASS. 2013. https://uwaterloo.ca/compass-system/compass-system-projects/past-projects/compass-cihr
- Hunter S., Veerasingam E., Barnett T.A., Patte K.A., Leatherdale S.T., Pabayo R. The association between income inequality and adolescent body mass index: Findings from the COMPASS study (2016–2019) Canadian Journal of Public Health. 2023:1006–1015. doi: 10.17269/s41997-023-00798-x.
- Jenson J., Saint-Martin D. New routes to social cohesion? Citizenship and the social investment state. The Canadian Journal of Sociology/Cahiers Canadiens de Sociologie. 2003;28(1):77–99. doi: 10.2307/3341876.
- Johnson J.L., Greaves L., Repta R. Better science with sex and gender: Facilitating the use of a sex and gender-based analysis in health research. International Journal for Equity in Health. 2009;8(1):14. doi: 10.1186/1475-9276-8-14.
- Kawachi I., Kennedy B. Income inequality and health: Pathways and mechanisms. Health Services Research. 1999;34(1 Pt 2):215–227.
- Kawachi I., Subramanian S.V. In: Social epidemiology. Berkman L.F., Kawachi I., Glymour M.M., editors. Oxford University Press; 2014. Income inequality; pp. 126–152.
- Khanolkar A.R., Patalay P. Socioeconomic inequalities in co-morbidity of overweight, obesity and mental ill-health from adolescence to mid-adulthood in two national birth cohort studies. The Lancet Regional Health - Europe. 2021;6 doi: 10.1016/j.lanepe.2021.100106.
- Leatherdale S.T., Brown K.S., Carson V., Childs R.A., Dubin J.A., Elliott S.J., Faulkner G., Hammond D., Manske S., Sabiston C.M., Laxer R.E., Bredin C., Thompson-Haile A. The COMPASS study: A longitudinal hierarchical research platform for evaluating natural experiments related to changes in school-level programs, policies and built environment resources. BMC Public Health. 2014;14(1):331. doi: 10.1186/1471-2458-14-331.
- Leatherdale S.T., Laxer R.E. Reliability and validity of the weight status and dietary intake measures in the COMPASS questionnaire: Are the self-reported measures of body mass index (BMI) and Canada's food guide servings robust? International Journal of Behavioral Nutrition and Physical Activity. 2013;10:1–11. doi: 10.1186/1479-5868-10-42.
- Li L. Gender differences in the relationship between income inequality and health in China: Evidence from the China Health and Nutrition Survey data. SSM - Population Health. 2024;25 doi: 10.1016/j.ssmph.2024.101601.
- Lin L., Bai S., Qin K., King C., Wong H., Wu T., Chen D., Lu C., Chen W., Guo V.Y. Comorbid depression and obesity , and its transition on the risk of functional disability among middle - aged and older Chinese : A cohort study. BMC Geriatrics. 2022:1–10. doi: 10.1186/s12877-022-02972-1.
- Lowe S.A.J., Hunter S., Patte K.A., Leatherdale S.T., Pabayo R. Exploring the longitudinal associations between census division income inequality and BMI trajectories among Canadian adolescent: Is gender an effect modifier? SSM - Population Health. 2023;24 doi: 10.1016/j.ssmph.2023.101519.
- Luppino F.S., de Wit L.M., Bouvy P.F., Stijnen T., Cuijpers P., Penninx B.W.J.H., Zitman F.G. Overweight, obesity, and depression: A systematic review and meta-analysis of longitudinal studies. Archives of General Psychiatry. 2010;67(3):220–229. doi: 10.1001/archgenpsychiatry.2010.2.
- Lynch J., Smith G.D., Harper S., Hillemeier M., Ross N., Kaplan G.A., Wolfson M. Is income inequality a determinant of population health? Part 1. A systematic review. The Milbank Quarterly. 2004;82(1):5–99. doi: 10.1111/j.0887-378x.2004.00302.x.
- McNeely C.A., Nonnemaker J.M., Blum R.W. Promoting school connectedness: Evidence from the national longitudinal study of adolescent health. Journal of School Health. 2002;72(4):138–146. doi: 10.1111/j.1746-1561.2002.tb06533.x.
- Melton P.A., Sims O.T., Oh H., Truong D.N., Atim K., Simon C. African American ethnicity, hypertension, diabetes, and arthritis independently predict Co-occurring depression and obesity among community-dwelling older adult alabamians. Social Work in Public Health. 2021;36(3):344–353. doi: 10.1080/19371918.2021.1895019.
- Milaneschi Y., Simmons W.K., van Rossum E.F.C., Penninx B.W. Depression and obesity: Evidence of shared biological mechanisms. Molecular Psychiatry. 2019;24(1):18–33. doi: 10.1038/s41380-018-0017-5.
- Mooreville M., Shomaker L.B., Reina S.A., Hannallah L.M., Adelyn Cohen L., Courville A.B., Kozlosky M., Brady S.M., Condarco T., Yanovski S.Z., Tanofsky-Kraff M., Yanovski J.A. Depressive symptoms and observed eating in youth. Appetite. 2014;75:141–149. doi: 10.1016/j.appet.2013.12.024.
- Muthén B., Asparouhov T., Sobel M.E. Applications of causally defined direct and indirect effects in mediation analysis using SEM in Mplus. 2011. https://api.semanticscholar.org/CorpusID:12409488
- Nasir N.S., Jones A., Mclaughlin M.W. School connectedness for students in low-income urban high schools. Teachers College Record: The Voice of Scholarship in Education. 2011;113:1755–1793. https://api.semanticscholar.org/CorpusID:140511533
- Pabayo R., Dunn E.C., Gilman S.E., Kawachi I., Molnar B.E. Income inequality within urban settings and depressive symptoms among adolescents. Journal of Epidemiology & Community Health. 2016;70(10):997–1003. doi: 10.1136/jech-2015-206613.
- Patalay P., Gage S.H. Changes in millennial adolescent mental health and health-related behaviours over 10 years: A population cohort comparison study. International Journal of Epidemiology. 2019;48(5):1650–1664. doi: 10.1093/ije/dyz006.
- Patel V., Burns J.K., Dhingra M., Tarver L., Kohrt B.A., Lund C. Income inequality and depression: A systematic review and meta-analysis of the association and a scoping review of mechanisms. World Psychiatry: Official Journal of the World Psychiatric Association (WPA) 2018;17(1):76–89. doi: 10.1002/wps.20492.
- Patte K.A., Gohari M.R., Leatherdale S.T. Does school connectedness differ by student ethnicity? A latent class analysis among Canadian youth. Multicultural Education Review. 2021;13(1):64–84. doi: 10.1080/2005615X.2021.1890310.
- Pei F., Wang Y., Wu Q., Shockley McCarthy K., Wu S. The roles of neighborhood social cohesion, peer substance use, and adolescent depression in adolescent substance use. Children and Youth Services Review. 2020;112 doi: 10.1016/j.childyouth.2020.104931.
- Pop I.A., van Ingen E., van Oorschot W. Inequality, wealth and health: Is decreasing income inequality the key to create healthier societies? Social Indicators Research. 2013;113(3):1025–1043. doi: 10.1007/s11205-012-0125-6.
- Prag P., Mills M., Wittek R. Income and income inequality as social determinants of health: Do social comparisons play a role? European Sociological Review. 2013;30:218–229. doi: 10.1093/esr/jct035.
- Preiss K., Brennan L., Clarke D. A systematic review of variables associated with the relationship between obesity and depression. Obesity Reviews: An Official Journal of the International Association for the Study of Obesity. 2013;14(11):906–918. doi: 10.1111/obr.12052.
- Qin K., Bai S., Chen W., Li J., Guo V.Y. Association of comorbid depression and obesity with cardiometabolic multimorbidity among middle-aged and older Chinese adults: A cohort study. Archives of Gerontology and Geriatrics. 2023;107 doi: 10.1016/j.archger.2022.104912.
- Quader Z.S., Gazmararian J.A., Suglia S.F. The relationships between childhood bullying, school connectedness, and adolescent adiposity, the fragile families child and wellbeing study. Journal of School Health. 2022;92(4):368–375. doi: 10.1111/josh.13138.
- Rajan T.M., Menon V. Psychiatric disorders and obesity: A review of association studies. Journal of Postgraduate Medicine. 2017;63(3):182–190. doi: 10.4103/jpgm.JPGM_712_16.
- Santini Z.I., Jose P.E., York Cornwell E., Koyanagi A., Nielsen L., Hinrichsen C., Meilstrup C., Madsen K.R., Koushede V. Social disconnectedness, perceived isolation, and symptoms of depression and anxiety among older Americans (NSHAP): A longitudinal mediation analysis. The Lancet Public Health. 2020;5(1):e62–e70. doi: 10.1016/S2468-2667(19)30230-0.
- Scott K.M., Bruffaerts R., Simon G.E., Alonso J., Angermeyer M., de Girolamo G., Demyttenaere K., Gasquet I., Haro J.M., Karam E., Kessler R.C., Levinson D., Medina Mora M.E., Oakley Browne M.A., Ormel J., Villa J.P., Uda H., Von Korff M. Obesity and mental disorders in the general population: Results from the world mental health surveys. International Journal of Obesity. 2008;32(1):192–200. doi: 10.1038/sj.ijo.0803701. 2005.
- Statistics Canada Census division: Detailed definition. 2018. https://www150.statcan.gc.ca/n1/pub/92-195-x/2011001/geo/cd-dr/def-eng.htm
- Valderas J.M., Starfield B., Sibbald B., Salisbury C., Roland M. Defining comorbidity: Implications for understanding health and health services. The Annals of Family Medicine. 2009;7(4):357–363. doi: 10.1370/afm.983.
- Van Dam N.T., Earleywine M. Validation of the center for epidemiologic studies depression scale--revised (CESD-R): Pragmatic depression assessment in the general population. Psychiatry Research. 2011;186(1):128–132. doi: 10.1016/j.psychres.2010.08.018.
- Viner R.M., Ozer E.M., Denny S., Marmot M., Resnick M., Fatusi A., Currie C. Adolescence and the social determinants of health. The Lancet. 2012;379(9826):1641–1652. doi: 10.1016/s0140-6736(12)60149-4.
- WHO BMI-for-age (5-19 years) 2007. https://www.who.int/tools/growth-reference-data-for-5to19-years/indicators/bmi-for-age
- WHO World health organization. AnthroPlus for personal computers. Manual: Software for assessing growth of the world ’ s children. 2007. http://www.who.int/growthref/tools/en/ Geneva.
- Wilkinson R.G. Health, hierarchy, and social anxiety. Annals of the New York Academy of Sciences. 1999;896:48–63. doi: 10.1111/j.1749-6632.1999.tb08104.x.
- Williams G.C., Patte K.A., Ferro M.A., Leatherdale S.T. Substance use classes and symptoms of anxiety and depression among Canadian secondary school students. Health Promotion and Chronic Disease Prevention in Canada. 2021;41(5):153–164. doi: 10.24095/hpcdp.41.5.02.
- Zahn-Waxler C., Klimes-Dougan B., Slattery M.J. Internalizing problems of childhood and adolescence: Prospects, pitfalls, and progress in understanding the development of anxiety and depression. Development and Psychopathology. 2000;12(3):443–466.
- Zhang Z., Wang Y., Zhao J. Longitudinal relationships between interparental conflict and adolescent depression: Moderating effects of school connectedness. Child Psychiatry and Human Development. 2023;54(5):1489–1498. doi: 10.1007/s10578-022-01355-2.
Associated Data
Supplementary Materials
Data Availability Statement
COMPASS study data is available upon request through completion and approval of an online form: https://uwaterloo.ca/compass-system/information-researchers/data-usage-application The datasets used during the current study are available from the corresponding author on reasonable request.