Mental Health Screening Outcomes Among Justice-Involved Youths Under Community Supervision

Abstract

The DISC Predictive Scales was administered to 812 New York City youths aged 10–19 placed under community supervision. Approximately half were indicated for possible mental health problems, most frequently mania and posttraumatic stress disorder. Girls were more likely than boys to flag on most disorders. Hierarchical Classes analysis produced five clusters: disruptive behavior, relational distress, marijuana, emotional dysregulation, and specific phobia. Posttraumatic stress disorder and mania were comorbid with all clusters except marijuana. Emotional dysregulation predicted higher, but relational distress predicted lower, risk for rearrest. Marijuana predicted failure to appear in court and receiving a final disposition of placement.

KEYWORDS:

• co-occurring disorders
• juvenile offenders
• mental illness

Notes

Note. Official DPS criteria: at least one symptom present and DPS impairment score equal to or greater than 6 OR total symptom score equal to or greater than 9 OR suicide attempt/ideation present OR any substance abuse present. Positive screen for specific DPS flag: Symptoms present and impairment score greater than or equal to 6.

*p < .05. **p < .01. ***p < .001.

^a N = 767.

^b N = 770.

*p < .05. **p < .01. ***p < .10.

Note. DB = disruptive behavior; RD = relational distress; MJ = marijuana; ED = emotional dysregulation; SP = specific phobia. Beta coefficients with SE B in parentheses. N = 746 for all models.

^a 0 = female, 1 = male.

*p < .05. **p < .01. ***p < .10.

Note. DPS = diagnostic predictive scales; DB = disruptive behavior; RD = relational distress; MJ = marijuana; ED = emotional dysregulation; SP = specific phobia. Column-wise cells within each cluster not sharing a subscript are different at p < .05 by Tukey's HSD test.

Note. DB = disruptive behavior; RD = relational distress; MJ = marijuana; ED = emotional dysregulation; SP = specific phobia. Step 1 variables were entered simultaneously. Step 2 variables were entered in a forward stepwise procedure with an entry criterion of p = .10. Rearrest indicators at 1 year from program intake. Fail to appear = three or more missed court appearances. All N = 773 except for Placement (N = 638).

^a 0 = female, 1 = male.

*p < .05. **p < .01. ⁺ p < .08.

There is another reason to conclude that Marijuana stands out as a unique cluster. HICLAS computes goodness-of-fit (GFI) coefficients for each participant that range from 0 to 1. These coefficients roughly correspond to factor loadings in that they indicate how well a given participant's actual data fit the prototypical pattern identified by the structure, with a value of 1 indicating perfect correspondence between the participant's actual flags and those predicted by the HICLAS model. Mean GFI was higher for the participants in the marijuana and specific phobia clusters than for disruptive behavior, relational distress, and emotional dysregulation (.72 and .71 vs. .59, .63, and .64), indicating that youth flagging for marijuana abuse more clearly constituted a “type” than a “fuzzy set.” Although the same is true for specific phobia, this cluster included only 10 noncomorbid participants (most of those linked to specific phobia were also linked to at least one other cluster).

The ADHD flag alone did not predict recidivism beyond the same set of control variables, beta = .27 (SE = 1.10), p > .35. Nor, in fact, did any of the other individual flags, ps > .08. The emotional dysregulation cluster of flags detected by HICLAS is thus a (somewhat) more valid indicator of recidivism risk than is any individual flag, including ADHD.