(PS4-B42) Exploring the Relationship Between Traumatic Stress Symptoms and Physical Restraint Incidents: A Parallel Mediation Analysis

Background: The early identification of youth at risk for physical restraint in Psychiatric Residential Treatment Facilities (PRTFs) is essential to reducing the likelihood of such incidents. Youth admitted to PRTFs report extensive histories of prior traumatization, and physical restraint may contribute to their retraumatization within the therapeutic milieu (Collin-Vézina et al., 2011; Roy et al., 2021). Notably, recent research indicates that youth who report the highest levels of traumatic stress symptoms at intake are the most likely to experience frequent physical restraint during their residential stay (Brown et al., 2022). Given the vicious circle between traumatic stress symptoms and physical restraint incidence, further exploration of the putative mechanisms of this relationship is warranted.

Method: The current study investigated potential mediators of the established relationship between traumatic stress symptoms and the frequency of physical restraint incidents. Participants included children (N = 150; 55% Female; 67% White, 33% Black or Biracial) ages 6-17 (M = 11.8 years) admitted to a PRTF in the northeastern United States. Variables significantly correlated (p < .05) with both traumatic stress symptoms and number of restraints were identified as potential mediators. All mediation models were tested with Hayes’ PROCESS macro (model 4) for IBM SPSS Statistics (Version 28.0).

Results: Potential mediators included emotional abuse severity, neglect severity, and symptoms of psychosis. Length of residential stay was included as a covariate within all analyses. Simple mediation models for neglect severity (ab = .34, SE = .15, 95% CI = [.06, .66]; c’ = 1.20, SE = .58, t = 1.97, p = .038, 95% CI = [.07, 2.34]) and psychotic symptoms (ab = .42, SE = .27, 95% CI = [.003, 1.04]; c’ = 1.15, SE = .58, t = 2.09, p = .051, 95% CI = [-.005, 2.31]) indicated partial and full mediation of the relationship between traumatic stress symptoms and restraint incidents, respectively. Emotional abuse severity was not a significant mediator. A parallel mediation model with neglect severity and psychotic symptoms as mediators was significant (F [2, 145] = 15.76, p < .001, R² = .18), and indicated full mediation of the relationship between traumatic stress symptoms and restraint incidents (ab = .80, SE = .36, 95% CI = [.21, 1.60]; c’ = .76, SE = .61, t = 1.24, p = .21, 95% CI = [-.45, 1.98]). Both psychotic symptoms (ab^a = .43, SE = .27, 95% CI = [.02, 1.06]) and neglect severity (ab^b = .37, SE = .17, 95% CI = [.07, .73]) were significant contributors to the observed effect.

 


Discussion: Results from the current study suggest that the severity of psychotic symptoms and prior neglect are critical indicators of risk for physical restraint among trauma-exposed youth in PRTFs. Comprehensive, trauma-focused assessment of these variables at intake in PRTFs may increase sensitivity of the identification of high-risk residents and the subsequent formulation of individualized treatment plans to target restraint reduction.