All clinical covariates were significantly associated with self-harm variables (see Supplemental Table 2 for tetrachoric correlations and Chi-Square test results), with the exception of depressed mood and having endorsed only one risky behavior. Significant correlations ranged from .09 (two risky behaviors) to .38 (anhedonia, regardless of depressed mood) for NSSI, and from .13 (two risky behaviors) to .46 (childhood sexual abuse) for SA. Females reported significantly more SA relative to males, χ2 (1, 6033)=7.23, p<.01, but there were no significant gender differences for NSSI. For all clinical covariates except the dummy coded variable encompassing individuals reporting only risky behavior, prevalence rates were higher among those with a history of self-harm (i.e., 95% confidence limits did not overlap). Again, this was not the case for depressed mood, which was removed from subsequent analyses due to the lack of an association with the self-harm variables1. The ORs between EAU and self-harm remained significant when adjusting for these clinical covariates (NSSI: OR=2.71, 95% CI: 2.10-3.51; SA: OR=2.56, 95% CI: 1.89-3.54; see Supplemental Table 1 for parameter estimates and test statistics).
