Descriptive statistics are presented in Table S1. 10,039 participants had the information for ACEs at T0 (mean ± SD: 2.08 ± 1.42). There were 474 (4.55%), 178 (1.63%), 412 (3.92%), and 65 (6.08%) participants who reported discrimination against racial/ethnic/color, non-US-born (non-United States-born) individuals, gay/lesbian/bisexual, and weight during the last 12 months at T1, respectively. The mean of the total score of the Perceived Discrimination Scale assessed at T1 was 1.18. In addition, 1,444 (14.77%) individuals reported lifetime experiences of perceived discrimination at T1. There were 1,216 (11.20%), 165 (1.52%), 845 (7.78%), and 462 (4.25%) adolescents reporting anhedonia, suicide attempt, suicide ideation and NSSI at T2, respectively.
Association between ACEs and discrimination
As shown in Fig. 1, after adjusting for age, sex, race/ethnicity, education, site, other subtype discrimination, ADI, family conflict, parental monitoring, school disengagement, school involvement, and school environment, ACEs were robustly associated with an elevated risk for discrimination against race/ethnicity/color (Odds ratio (OR) = 1.23, 95% CI = 1.12, 1.36, p < 0.001), discrimination against gay/lesbian/bisexual (OR = 1.28, 95% CI = 1.16, 1.41, p < 0.001), discrimination against weight (OR = 1.12, 95% CI = 1.03, 1.21, p < 0.001), and perceived lifetime discrimination (OR = 1.10, 95% CI = 1.03, 1.16, p = 0.002) (Table S3).
Association between ACEs, discrimination and anhedonia
Figure 2 illustrates that ACEs (OR = 1.13, 95% CI = 1.06, 1.21, p < 0.001), discrimination against gay/lesbian/bisexual (OR = 1.53, 95% CI = 1.02, 2.31, p = 0.040), discrimination against weight (OR = 1.47, 95% CI = 1.05, 2.05, p = 0.025) and lifetime experiences of perceived discrimination (OR = 1.48, 95% CI = 1.16, 1.88, p = 0.002) were significantly associated with anhedonia, when controlling age, sex, race/ethnicity, education, site, ADI, family conflict, parental monitoring, school disengagement, school involvement, and school environment (Table S4).
Association between ACEs, discrimination, anhedonia and suicidality
In Fig. 3, ACEs (OR = 1.17, 95% CI = 1.08, 1.26, p < 0.001), discrimination against gay/lesbian/bisexual (OR = 2.22, 95% CI = 1.46, 3.35, p < 0.001), discrimination against weight (OR = 2.01, 95% CI = 1.42, 2.84, p < 0.001), lifetime experiences of perceived discrimination (OR = 1.38, 95% CI = 1.04, 1.83, p = 0.024) and anhedonia (OR = 3.25, 95% CI = 2.54, 4.15, p < 0.001) had significant association with an increased risk for suicidality, after adjusting the covariates (Table S5).
Association between ACEs, discrimination, anhedonia and NSSI
In Fig. 4, ACEs (OR = 1.14, 95% CI = 1.03, 1.26, p < 0.001), discrimination against gay/lesbian/bisexual (OR = 2.99, 95% CI = 1.86, 4.81, p < 0.001), lifetime experiences of discrimination (OR = 1.53, 95% CI = 1.07, 2.18, p = 0.019), and anhedonia (OR = 3.31, 95% CI = 2.44, 4.50, p < 0.001) had significant positive effects on NSSI, after adjustment of the covariates (Table S6).
Serial mediation model of discrimination against gay/lesbian/bisexual and anhedonia in the association between ACEs and outcomes
Serial mediation analyses were conducted to examine the sequentially mediating effects of discrimination against gay/lesbian/bisexual (T1) and anhedonia (T2) on the association between ACEs (T0) and outcomes of suicidality and NSSI (T2), respectively. In Fig. 5A, after adjusting for covariates, the total effect of ACEs on suicidality was significant (β = 0.117, SE = 0.024, 95% CI = 0.042, 0.087). Specifically, the indirect effect of ACEs on suicidality, mediated by discrimination against gay/lesbian/bisexual and anhedonia, was 0.027 (SE = 0.010, p = 0.008) and 0.020 (SE = 0.008, 95% CI = 0.042, 0.087) respectively. The serial indirect effect of the pathways of ACEs → discrimination→ anhedonia → Suicidality was estimated to be 0.006 (SE = 0.003, 95% CI = 0.042, 0.087). The total mediation effect was significant (β = 0.054, SE = 0.012, 95% CI = 0.042, 0.087), indicating that the relationship between ACEs and suicidality was partially mediated by the mediators (proportion mediated: 46.15%). The full model explained 22.2% of the variance in suicidality (please see Table S7 and Table S8 for more details). In Fig. 5B, after the covariates were adjusted, it was observed that the direct effect of ACEs on NSSI was not significant (β = 0.039, SE = 0.012, 95% CI = -0.012, 0.093). However, the indirect mediation effect was significant (β = 0.062, SE = 0.012, 95% CI = 0.042, 0.087) (proportion mediated: 61.89%). The pathways (ACEs → discrimination→ NSSI and ACEs → anhedonia → NSSI) were all significant, with the estimate of 0.037 (SE = 0.012, 95% CI = 0.042, 0.087) and 0.019 (SE = 0.007, 95% CI = 0.042, 0.087), respectively. The serial indirect effects of the pathways of ACEs → discrimination→ anhedonia → NSSI was estimated to be 0.006 (SE = 0.003, 95% CI = 0.042, 0.087). The full model accounted for 22.7% of the variance in NSSI (Table S7 and S8).