(PS2-D88) Emotion Reactivity and Substance Use as Independent and Interacting Risk Factors for Nonsuicidal Self-injury

Greater levels of emotion reactivity (ER) – or how intensely and for what duration one experiences an emotion – have been associated with increased risk for nonsuicidal self-injury (NSSI; Nock et al., 2008). ER is thought to be a predisposing factor for deficits in emotion and behavior regulation (Nock et al., 2008), and thus ER may also increase an individual’s tendency to engage in impulsive, risky, and self-injurious behavior more generally. The frequent co-occurrence of NSSI and substance use disorders (SUDs; Moller et al., 2013; Nock et al., 2006) has been discussed in relation to emotion dysregulation (e.g., Gratz & Tull, 2010; Nock, 2010), though ER has been less frequently studied in this context. The present study examined the extent to which the presence of an alcohol or other drug SUD moderated the relationship between ER and NSSI history and frequency in a large collegiate sample (n = 8,065). We hypothesized that both ER and SUD would be significant predictors of NSSI engagement, and that the ER-NSSI relationship would be stronger for people with an SUD. Participants completed self-report measures of NSSI (Form And Function of Self Injury Scale; Jenkins, 2011) and ER (Emotion Reactivity Scale [ERS]; Nock et al., 2008), as well as the Alcohol Use Disorder and (other) Drug Use Disorder Identification Tests (AUDIT and DUDIT; Saunders et al., 1993 & Babor et al., 2011, respectively). The AUDIT and DUDIT were scored and interpreted using the cutoff score that indicates likely SUD (Saunders et al., 1993). A logistic regression model revealed that a positive DUDIT screen and higher ERS total scores were predictive of having engaged in NSSI (DUDIT: OR = 2.137, ERS: OR = 1.033, ps < .001), while a positive AUDIT screen was not (p = .986). These results were moderated by a significant DUDIT x ERS interaction (OR = 1.007, p = .044). Consistent with our hypothesis, simple slope analyses revealed that, though ERS score was associated with NSSI in both groups, the association was stronger among those with a positive DUDIT screen. There was no significant AUDIT x ERS interaction (p = .827). Among only those participants with a history of NSSI (n = 2,575), a negative binomial regression demonstrated that the DUDIT and ERS were positively associated with number of lifetime NSSI acts (bs = 0.444 & 0.009, respectively, ps < .001). In contrast, AUDIT was negatively associated with lifetime NSSI acts (b = -0.342, p < .001). These main effects were moderated by DUDIT x ERS (b = 0.022, p < .001) and AUDIT x ERS (b = 0.008, p = .024) interactions. Simple slope analyses again demonstrated that the association between ERS and NSSI frequency was stronger among participants with positive DUDIT or AUDIT screens as compared to those without. Taken together, our findings suggest that (non-alcohol) SUD and ER are independent predictors of NSSI engagement and frequency that also interact to confer additive risk.