Although coercive physical interventions such as restraint and seclusion have been linked to serious injuries and death in children and adolescents, little is known about the risk factors for youths' involvement in these interventions, particularly in the context of long-term residential treatment. This study examined admission and intake records for 99 youths in an intensive residential treatment facility to identify risk factors that could predict involvement in coercive physical interventions, to identify patterns of coercive intervention involvement while in treatment, and to explore the role of staff training in the use of these interventions. Results indicated that younger age and higher ratings of impulsivity predicted both any involvement and the total involvement in coercive interventions. While aggression was positively correlated with any involvement in coercive interventions, it did not predict total number of interventions required. Higher ratings of callous and unemotional traits and longer length of stay also predicted the total number of interventions a youth was involved in. Five trajectories of involvement across the first eight months of treatment were also identified, including a minimally involved group, a low-slightly increasing group, a moderate-increasing group, a moderate decreasing group, and a high-decreasing group. Finally, results indicated that staff members appeared more likely to utilize coercive physical interventions immediately after training, with steadily declining utilization until the next training session. Also discussed are the relevance of these results to residential treatment and the next steps for research on coercive physical interventions, given that these youths frequently interact with multiple interconnected microsystems, each of which exerts influence on youths' responses to treatment.