Effects of a Randomized Clinical Trial to Enhance Risk Detection Among Homeless Youth

Background: The majority of homeless youth experience victimization before leaving home and encounter subsequent victimization on the streets, including robbery, physical attacks, and sexual assault. Previous research indicates youth with histories of trauma often fail to detect danger risks, making them vulnerable to subsequent victimization. The current study tested a mindfulness-based, cognitive, skill-building intervention designed to train homeless youth to better detect risk through focusing attention to internal cues (physiological responses, cognitions), interpersonal cues (controlling, suspicious behavior by others) and environmental cues (dark, isolated, unfamiliar situations). The study thus investigated: (1) whether the intervention was successful in increasing homeless youths’ overall risk detection abilities, and, if so, (2) which specific types of risk detection were positively affected.

Methods: Youth ages 18-24 (N=84) were recruited from a homeless youth shelter and randomly assigned to receive usual case management services (goal-setting and resource referral) or usual services plus a 3-day manualized intensive risk detection intervention. Pretest and posttest (1 week later) interviews assessed youths’ risk detection abilities through a series of vignettes, read aloud to youth, describing characters in risky situations, and asking youth to identify risk cues present. Youths’ open-ended responses were coded by a research team utilizing a standardized codebook (93.4% inter-rater reliability). Researchers coded each identified cue as an internal, interpersonal, or environmental cue before calculating the proportion of cues identified out of total cues embedded in each vignette. Univariate analyses compared youths’ abilities to detect different forms of risk (internal, interpersonal, and environmental) prior to intervention. Bivariate analyses tested whether the intervention group faired better than the control group in overall risk detection and in their abilities to detect specific risk types (internal, interpersonal, and environmental).

Results: On average, youth at pretest identified 23% (M=.23, SD=.10) of risk cues and were able to identify a greater proportion of interpersonal cues (M=.34, SD=.23) compared to environmental cues (M=.20, SD=.24) and internal cues (M=.13, SD=.14). While the intervention and control groups did not differ significantly at pretest, by posttest, youth in the intervention group scored significantly higher (M=.32, SD=.13) than controls (M=.26, SD=.13) on overall risk detection (t(76)=-2.04, p<.05). Regarding specific types of risk detection, intervention youth performed better (M=.49, SD=.25) than controls (M=.29, SD=.25) in identifying interpersonal risk cues (t(76)=-3.28, p<.05). Although the trend in identifying environmental cues favored the intervention group (M=.29, SD=.30) compared to control (M=.24, SD=.23), it did not reach statistical significance. Groups were similar in their abilities to identify internal risk cues (t(76)=.55, p=.57).

Conclusions: This risk detection intervention appears promising in improving homeless youths’ abilities to detect risk on the streets. However, youths’ abilities to recognize when others are acting suspicious or dangerous (interpersonal cues) appear much more malleable compared to youths’ abilities to recognize environmental or internal cues that may require awareness of subtle situations and physiological reactions. Further iterations of this intervention should work to improve attention to internal danger cues skills, as these were the most difficult to change and the least commonly-identified among youth at the start of the trial.