Association of Social Support During Adolescence With Depression, Anxiety, and Suicidal Ideation in Young Adults

This cohort study assesses whether perceived social support during adolescence is associated with less depression, anxiety, and suicidal ideation and fewer suicide attempts in young adults.


Introduction
Perceived social support refers to the subjective availability of care and assistance received from social relationships and it is characterized by emotional support (eg, expression of empathy), instrumental support (eg, help with household tasks) and informational support (eg, financial advice) that can be provided from various sources, such as friends or family. 1 Findings of systematic reviews 2,3 and a meta-analysis 4 have suggested that individuals who perceive greater social support experience fewer depressive symptoms. However, most of the evidence is based on cross-sectional studies that did not control for prior mental health problems (MHPs), providing no insight into directionality of the association between perceived social support and MHPs. This information gap is problematic because individuals who previously experienced MHPs may be more likely to perceive a lower level of social support and to have smaller social networks of lesser quality, resulting in inflated effect sizes. 5,6 In addition, many studies use social support measures that lack strong psychometric properties. 2,6 Social support may be especially important during emerging adulthood (late teens to late 20s), a period simultaneously characterized by major changes in social roles and responsibilities (eg, living independently, transitioning from mandatory education to college or the workforce, establishing romantic relationships 7 ) and increased incidence of MHPs. 8,9 Importantly, MHPs in youth may be associated with lifelong consequences such as lower educational attainment, unemployment, and chronic MHPs. [10][11][12][13] The evidence regarding associations between perceived social support and MHPs in emerging adulthood remains limited. Few longitudinal studies have examined the role of social support on MHPs in emerging adulthood, while accounting for baseline and/or previous MHPs. 14 -16 In 2 longitudinal studies, social support from a non-specified source was found to be protective against depression. 14,15 However, parental social support was found to be protective against depression only when instilled earlier on in development. 17,18 Evidence remains scarce for other salient and clinically significant MHPs, including generalized anxiety 17 and suicidal ideation, 16 with no studies examining the protective role of social support for suicide attempts during emerging adulthood. Because well-established sex-based differences exist across MHPs, it is imperative to investigate potential sex-based differences in perceptions of social support and the extent to which male and female individuals may benefit differently from social support.
Using a large population-based cohort from the Canadian province of Québec, we aimed to assess whether individuals who perceive higher levels of social support (from a non-specified source) in emerging adulthood have a reduced risk for later MHPs, including depression, anxiety and suicidal ideation and attempts, 1 year later. Because emerging adults may have already exhibited MHPs prior to this transitional life stage, we accounted for a range of MHPs during adolescence and family characteristics. To our knowledge, this study is the first to investigate the association of perceived social support with a range of common MHPs simultaneously within a contemporary populationbased cohort of emerging adults (aged 19-20 years) while taking into account previous MHPs and family characteristics to clarify the directionality of these associations.

Method Participants
The Quebec Longitudinal Study of Child Development (QLSCD) is an ongoing population-based cohort that includes 2120 participants born from 1997 through 1998 in Quebec. The QLSCD is conducted by the Institut de la Statistique du Québec. 19 This cohort study was approved by ethical committees of the Québec Institute of Statistics of the CHU Sainte-Justine Hospital Research Centre and participant written informed consent was obtained. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. 20 Of the 1680 individuals contacted at the 20-year data collection point, 1235 responded, and 1174 provided information on social support at age 19 years and MHPs at age 20 years (56% of the original 2120 participants). Similar to previous studies that used this cohort, we selected comparison variables that have the potential to identify the most vulnerable participants and, in turn, those most likely to be lost at follow-up. 21,22 Participants who were underrepresented were more likely to be male, to be of non-Canadian ancestry, to have a younger mother at birth and/or with depressive symptoms, to come from a nonintact family, to have parents with a low socioeconomic status at age 5 months, and to have higher externalizing symptoms at age 29 months (eg, cannot sit still or is agitated) as measured by 10 items from the Behavior Questionnaire, 23 with scores ranging from 0 to 18 and high scores indicating higher levels of externalizing symptoms (eTable 1 in the Supplement).
To minimize bias attributable to such differential attrition, we applied inverse probability weighting based on variables conditional to attrition in all analyses.

Measures Perceived Social Support at Age 19 Years
Perceived social support was assessed using the 10-item Social Provision Scale (SPS-10 24 ), a shortened version of the original 24-item scale, 25 which measures dimensions of attachment, social integration, reassurance and worth, reliable alliance and guidance based on statements (eg, "there is someone I can talk to about important decisions in my life") rated on a Likert scale (range, 0 = strongly disagree to 4 = strongly agree). The SPS-10 is a valid and reliable measure of perceived availability of social support (Cronbach α = .89). 24,26 All SPS-10 scores were converted to z scores, with higher scores indicating higher levels social support.

Mental Health Problems at Age 20 Years
Depressive symptoms experienced during the past week were assessed using the Centre for Epidemiological Studies Depression scale (CES-D) short-form, 27 including 12 statements 28,29 (eg, "I felt depressed") rated on a Likert scale from 0 = rarely/none of the time to 3 = most/all of the time.
Scores ranged from 0 to 36, with higher scores representing increasing symptom severity. The scale has good internal reliability (Cronbach α = .85). Anxiety symptoms experienced during the past 2 weeks were assessed using the Generalized Anxiety Disorder 7-item scale 30 (GAD-7), with statements (eg, "Feeling nervous, anxious or on edge") rated from 0 = not at all to 3 = nearly every day. Scores range from 0 to 21, with higher scores representing increasing symptom severity. The GAD-7 is a reliable measure of anxiety symptoms (Cronbach α = .89).
Serious suicidal ideation was assessed with the following question: "In the past 12 months, did you ever seriously think of attempting suicide?" If the answer was affirmative, they were asked how many times they attempted suicide in the last 12 months. Both of these variables were dichotomized as 0 or Ն1. 31

Adolescent Confounders
We measured confounders at ages 15 and 17 years and averaged the scores across ages, unless otherwise indicated. Several factors were identified as having associations with both the MHPs of interest (depressive and anxiety symptoms and suicidal ideation and attempts) and perceived social support (eTable 2 in the Supplement). These factors included sex; family socioeconomic status assessed as an aggregate of parental occupation, parental education level, and annual gross income 32 ; family functioning (ie, communication, problem resolution, and control of disruptive behavior assessed using the 7-item McMaster Family Assessment Device 33 [scores range from 0-10, with higher scores indicating more family dysfunction]); and family structure at age 17 years (biological parents, blended family, single parent), as well as MHPs in the last 12 months (measured using the Mental Health and Social Inadaptation Assessment 34 ), including for depressive, generalized anxiety, oppositional and/or defiant, inattention or hyperactivity, and conduct disorder symptoms.
Subscales for these symptoms are scored as follows: 0 = never true, 1 = sometimes true, 2 = always true. Higher scores represent increasing symptom severity. Suicidal ideation and suicide attempts were measured using the same instruments at age 20 years.

Statistical Analyses
All statistical analyses were conducted using IBM SPSS, version 26. Prospective associations between perceived social support and MHPs in emerging adulthood were estimated using linear or binary logistic regressions. The first model was adjusted for sex only, whereas the second model was additionally adjusted for adolescent confounders, including MHPs (depressive, anxiety, oppositional/ defiant, inattention/hyperactivity and conduct disorder symptoms and suicidal ideation or attempts) and family characteristics (socioeconomic status, family functioning and structure). No sex-based differences were found in social support interactions across prospective associations; thus, we present regression results with sexes combined. Missing data for confounders ranges from 6% (MHPs) to 18.3% (family structure at 17 years). To avoid further loss of data, missing data on confounders were imputed using multiple imputation by chained equations, 35 and analyses were conducted across 100 imputed data sets (N = 1174). Patterns of results based on multiple imputations were comparable to those based on maximum available cases (eTables 3 and 4 in the Supplement).
Using multinomial logistic regressions, we further estimated the clinical significance of prospective associations between social support and depressive and anxiety symptoms by taking into consideration symptom severity based on validated cutoff values on the CES-D and GAD-7 scales. No or minimal symptoms had scores ranging from 0 to 11 for depression and from 0 to 9 for anxiety on the CES-D and GAD-7 scales, respectively, moderate symptoms had scores ranging from 12 to 20 for depression and 10 to 14 for anxiety, and severe symptoms had scores ranging from 21 to 36 for depression and 15 to 21 for anxiety. 29,30

Results
The total sample consisted of 1174 participants, including 574 female (48.89%) and 600 male (51.11%) individuals. Social support was measured at age 19 years and mental health outcomes were measured at age 20 years. The descriptive statistics on key variables are presented in Table 1. Overall, male individuals perceived significantly less social support than female individuals at age 19 years.
Mental health problems were assessed at age 20 years, and of the total sample, 289 individuals (24.6%) reported moderate depressive symptoms and 67 (5.7%) reported severe symptoms. One  Table 1). No sex-based differences were observed for suicide-related outcomes. Additionally, correlations between adolescent confounders and the key variables including social support at 19 years and MHPs at age 20 years are described in eTable 2 in the Supplement. Of note, all MHPs at ages 15 and 17 years, except suicide attempts, were significantly correlated with social support at age 19 years (Pearson r range, −0.10 to −0.20; P < .05) and MHPs at age 20 years (Pearson r range, −0.08 to 0.45; P < .05).

Associations Between Social Support and MHPs
Prospective associations between perceived social support at age 19 years and MHPs at age 20 years, are shown in Table 2

Associations Between Social Support and Clinical Thresholds of MHPs
Individuals who perceived higher levels of social support were less likely to report moderate or severe depressive and anxiety symptoms compared with those with no or minimal symptoms in fully adjusted models ( Table 3). For example, every SD increase in social support lowered the odds for

Discussion
Using data from the QLSCD, we found that emerging adults who perceived a greater level of social support were less likely to report MHPs including suicidal ideation and attempts 1 year later, independent of previous MHPs and family characteristics. Benefits of social support were observed across the full spectrum of depressive and anxiety symptoms, including clinically increasing severe  including anxiety, depression, suicidal ideation and suicide attempts, attention deficit/hyperactivity-impulsivity, oppositional defiant and conduct symptoms as well as family characteristics in adolescence, including socioeconomic status, family functioning, and family structure (at age 17 years). Reference category: no or minimal symptoms for depressive and anxiety symptoms. symptoms after adjusting for baseline depressive symptoms. For instance, increases in social support were associated with shifts in depression trajectories. 15 Yet another study that adjusted for baseline depressive symptoms found that social support was beneficial only in youth facing work and financial stress. 14 However, a longitudinal study 18 failed to detect an association of social support with trajectories of depressive symptoms, but this inconsistency may have occurred because of the study's focus on parental support, which may be protective only when established earlier in life. 17 We found that perceived social support was associated with fewer generalized anxiety symptoms, including severe anxiety. This finding is consistent with a prospective study that found that lower perceived parental social support in childhood was associated with a 4-fold higher risk of generalized anxiety disorder in emerging adulthood. 17 However, previous MHPs were not accounted for and social support was based on retrospective reports. We believe that the findings of the present study support evidence from previous studies by showing that similar associations are observed between perceived social support and anxiety symptoms in emerging adulthood, while accounting for adolescent MHPs.
In addition, we found that social support was associated with less suicidal ideation in emerging adults. A previous study with female participants found that those who enjoyed higher levels of social support were more likely to be in remission from suicidal thoughts. 16 Social support was also associated with fewer suicide attempts in emerging adults. The protective benefits of parental social support against suicide attempts during adolescence has been demonstrated. 42 However, to our knowledge, this is the first study to highlight the association of perceived social support with fewer suicide attempts in emerging adulthood. A randomized clinical trial found that support from a Youth-Nominated Support Team (ie, characterized by psychoeducation and weekly emotional and instrumental support check-ins from supportive adults for 3 months) was associated with a reduced risk for suicide mortality in vulnerable adolescents aged 13 to 17 years. 43 The limited prospective evidence using suicide attempts as an outcome is consistent with that of a cross-sectional study 44 and points to the role of social support in suicide-related outcomes. The use of a social support measure with an unspecified source in the present study allowed respondents to reflect on a source of support available to them, rather than being restricted to reflect on a specific source. Results of this population-based study are also in line with those of previous studies reporting the protective benefits of social support for vulnerable groups (eg, adolescents who experience bullying, 45 clinical populations, 46 non-heterosexual populations 47 and those exposed to higher levels of stress 14 ).

Strengths and Limitations
This study has several strengths, including the prospective design and the ability to adjust for a range of confounders (previous MHPs and family characteristics), which addresses to the limitations of reverse causality found in previous studies. 2 Without adjusting for previous MHPs, we cannot effectively rule out the possibility that associations of social support with MHPs are better accounted for by the associations of MHPs with social support. Moreover, in contrast to several previous studies, 2,6 social support was measured using a validated tool with excellent psychometric properties. 24 Another major strength of the present study is that we investigated associations with clinically significant outcomes (ie, generalized anxiety and suicide-related outcomes) 48  sizes. However, in this population-based study, youth were the only informants given that they offer the most insight into their own suicidal behaviors 49 and internalizing symptoms in comparison to parents or other informants. Although this study does not measure objective reports of social support or MHPs, future studies may consider using other informant sources to ascertain whether social support in itself is protective for mental health (ie, the objective measure of support), or whether the benefits of social support lie in the perception of it. Third, our study did not control for shared genetic vulnerability between both social support and MHPs. 50 Although social support may be considered to be an environmental factor, research has pointed toward the role of genetic factors in social support. 51,52 Using the twin study design, 2 studies found that social support and depression in adolescence were phenotypically associated, yet such associations were largely explained by genetic factors affecting both the perception of social support and depression. In contrast, a study using a discordant twin design in the Environmental Risk Longitudinal Twin Study (E-RISK) cohort found that the association of social support with psychotic symptoms was largely environmentally mediated. 53 Thus, future studies should further investigate the magnitude of associations while considering a genetically informative design to untangle the unique advantages of social support.
Finally, causality cannot be established by observational studies alone because the possibility that associations are better explained by unmeasured confounders cannot be ruled out. However, the findings of the present study are consistent with those of a randomized clinical trial reporting the protective benefits of social support in suicide-related outcomes. 43

Conclusions
Emerging adulthood is a transitional life period marked by a high prevalence of MHPs. This study provides evidence on the benefits associated with social support for MHPs and suicide-related outcomes during this life-period, even in individuals who experienced MHPs in an earlier stage of development. Therefore, this study raises awareness of the potential protective role of perceived social support for mental health during emerging adulthood and provides evidence for the importance of leveraging social support in treatment options while taking into consideration perceptions of support.