A large body of research has focused on mental health characteristics as key risk factors for suicidality. Depression is one of the most well-studied risk factors for suicide risk; a recent meta-analysis examining predictors of suicidal ideation and attempts found that elevated depression predicted higher rates of ideation, but depression severity did not differ between those who ideated and those who attempted suicide [8]. Similar associations have also been observed in clinical samples of adolescents [9–11]. Prior research has also examined the relationship between disordered eating, panic disorder, and suicidality, demonstrating a strong relationship between suicide and eating disorder diagnoses as well as even sub-threshold eating disorder symptoms [12, 13]; unhealthy weight control behaviors, such as fasting and purging, have also been found to predict higher odds of suicidal ideation and attempts in adolescent samples [14, 15]. Conflicting findings have emerged from studies on panic disorder in adolescents, with some reporting no evidence of a relationship between panic disorder and suicidal ideation or attempts [16, 17], and others reporting significant associations even when controlling for other comorbid disorders [18, 19].

Connectedness in adolescent suicide research has been operationalized as self-reported feelings of being supported or cared for as well as social bonds in settings such as schools and communities [20]. Ideation-to-action theories, which frame the development of suicidal ideation and progression from ideating to attempting suicide as distinct processes, have identified both mental health and connectedness as key components in these processes. The interpersonal theory of suicide [5] proposes that feelings of sadness and hopelessness along with thwarted belongingness result in the development of suicidal ideation, whereas habituation to pain leads to the acquired capability to attempt suicide. In addition, the three-step theory [6] postulates: (1) the combination of hopelessness and psychological pain result in the development of moderate suicidal ideation; (2) moderate ideation escalates to strong ideation when psychological pain and hopelessness exceed connectedness; and (3) strong ideation progresses to attempts when increased capacity—facilitated by dispositional, acquired, and practical factors—is present. Empirical tests have supported these theories, highlighting the importance of increasing connectedness in individuals reporting mental health concerns [21, 22].

A recent meta-analysis found that stronger school connectedness was associated with reduced odds of suicidal ideation and attempts in community (i.e., general population) and higher-risk (e.g., sexual and gender minorities) samples of youth [23]. In particular, prior research on school connectedness suggests social connections with teachers [24–27] and peers [26, 28, 29] are associated with lower rates of suicidal ideation and/or attempts. Additionally, school safety indicators as measures of structural support have been identified as significant predictors of suicidality [30–33]. Although few studies have examined the relationship between suicidality and connectedness within broader social contexts, various measures of community connectedness have been associated with reduced ideation and/or attempt risk in adolescent samples [27, 32, 34, 35]. Furthermore, engagement in community service and helping activities may be beneficial as prior research has found that volunteerism [36] and perceived prosocial attributes [37] were associated with reduced suicide risk levels and ideation, respectively.

Youth spend a considerable amount of time at school and within their communities during adolescence, prompting the need to examine whether social connections forged in these extra-familial settings can help buffer the elevated risk of suicidality associated with mental health conditions. A prior study examining the relationship between mental health conditions and suicidal ideation in a sample of U.S. high school students found that adolescents with disordered eating and low school-based peer support reported more severe suicidal ideation than adolescents with disordered eating and high school-based peer support [28]. However, the same study found the relationship between depression levels and ideation severity did not vary across different levels of school-based peer support. Conversely, another study conducted with a community sample of adolescents found that community connectedness buffered the association between depression and reasons for living (a proxy for suicide risk), such that youth with more severe depression and high levels of community connectedness reported more reasons for living comparted to those with low or moderate community connectedness [38]. Although the buffering effects of volunteerism and community service has been understudied, one prior study found that the relationship between depression and suicidal ideation was buffered by prosocial attributes in samples of young adults in Europe and Asia [37]. Taken together, these findings suggest that school and community connectedness may buffer the association between certain mental health conditions and suicide risk, although results may vary across samples or types of support examined.

To the best of our knowledge, prior research examining these protective effects have been limited to community samples of youth, creating a gap in knowledge regarding whether mental health conditions and social connectedness interact to predict lower suicidal ideation and attempt risk in clinical samples of adolescents (i.e., those receiving mental or behavioral health treatment). Under a comprehensive public health approach to suicide prevention, it is important to better understand how social connections might be leveraged to buffer the effects of mental health conditions among youth most vulnerable to suicidal thoughts and behaviors. As such, the current study seeks to help fill this gap in knowledge by assessing: (a) which mental health conditions (depression severity, disordered eating, panic disorder) are associated with suicide risk; and (b) which measures of school connectedness (teacher support, school-based positive peer interactions, school safety) and community connectedness (neighborhood social connection and community engagement) are associated with suicide risk; and (c) whether these connectedness measures buffer the associations between mental health conditions and suicide risk.

The results of this study provide crucial knowledge about the relationships between mental health conditions, social connectedness, and suicide risk in a clinical sample of adolescents with histories of substance use disorders. Results indicated that MDD severity and panic disorder diagnosis were associated with elevated suicidal ideation risk, whereas disordered eating was associated with elevated risk of suicide attempts. We also found that higher levels of school-based positive peer interactions, school safety, and neighborhood social connection were associated with reductions in suicide attempt risk, even after adjusting for MDD severity. Moderation analyses indicated that adolescents who screened positive for disordered eating and reported low neighborhood social connection were at the greatest risk of attempting suicide, but at high levels of neighborhood social connection, their attempt risk was equivalent to those who screened negative for disordered eating.

To the best of our knowledge, this is the first study to examine the potential protective effects of school and community connectedness in a clinical sample of adolescents with histories of substance use disorders. As such, these findings fill an important gap as they suggest that similar clinical populations may benefit from approaches that aim to promote social connectedness, further supporting a comprehensive approach to suicide prevention [4]. Whereas the increased suicidality risk associated with MDD, disordered eating, and panic disorder highlight the importance of implementing evidence-based clinical treatments to address these mental health conditions, the reduced attempt risk associated with school-based positive peer interactions, school safety, and neighborhood social connection underscore the importance of efforts to bolster supportive relationships and environments. Given the finding that higher levels of neighborhood social connection may partially mitigate the elevated likelihood of attempting suicide associated with disordered eating, stronger community ties may be particularly beneficial for adolescents who engage in disordered eating behaviors.

The results of this study also revealed differential relationships between certain mental health conditions and suicide risk level. Consistent with results of prior studies conducted with other adolescent clinical samples [10, 11], the findings from both our unadjusted and adjusted models indicate that MDD severity is a risk factor for suicidal ideation but not attempts among youth with histories of substance use disorders. Similarly, despite some studies reporting no evidence of a relationship between panic disorder and suicide risk [16, 17], our results support prior findings of a significant relationship between panic disorder and ideation risk [18, 19]. In contrast, screening positive for disordered eating was associated with higher risk of suicidal ideation and attempts in unadjusted risk group comparisons, but only the association with increased suicide attempt risk remained significant after adjusting for covariates. A possible explanation for the unique relationship between disordered eating and increased attempt risk is that individuals who repeatedly engage in pain-inducing weight control behaviors (e.g., overexercising and purging) demonstrate higher acquired capacity for suicide due to greater pain insensitivity, making them more likely to attempt suicide than those who do not engage in those behaviors [47–49]. Moreover, the relationship between disordered eating and attempt risk may be further bolstered by high comorbidity with traumatic experiences. Approximately 70% of youth who screened positive for disordered eating also reported histories of physical abuse, sexual abuse, or being attacked with a weapon, all of which were associated with higher odds of attempting suicide [9].

We found no evidence that school or community connectedness buffered the associations between mental health conditions and suicidal ideation in this sample. Given that psychopathology and low social connectedness are thought to be key contributing factors to suicidal ideation [5, 6], we expected measures of school and community connectedness to be significantly lower in the ideation only group compared to the non-suicidal group. This unexpected result could be a measurement artifact given the tools used to measure suicidality. As suggested by the three-step theory [6], connectedness may differ between those with moderate (i.e., passively thinking about suicide) versus strong ideation (i.e., active desire to die by suicide suicide), but our ideation only group was based on a M.I.N.I. item that was a unidimensional assessment of ideation. Thus, any protective effects present may have gone undetected when comparing a potentially heterogeneous ideation only group to the non-suicidal group. In contrast, other studies [28, 37] that have reported significant buffering effects between mental health conditions and suicidal ideation used instruments that measured ideation on a semi-continuous scale, and as such, were able to assess ideation severity. It is also possible that these null findings can be attributed to ceiling effects resulting from our clinical sample of adolescents who have more severe mental health conditions than would be expected in a community sample. As such, this may account for the differences between our findings and studies conducted with community samples [28, 37, 38]. Finally, there may be other dimensions of social connectedness that are important to study that were not accounted for in the present study (e.g., familial connectedness), highlighting the need for future research to examine the effects of additional dimensions of social connectedness in other clinical samples of youth.