What are effective strategies for implementing trauma-informed care in youth inpatient psychiatric and residential treatment settings? A realist systematic review

After extracting and systematically analyzing data, we observed five main factors in our analysis of cross-site TIC implementation: (1) the critical importance of senior leaders prioritizing TIC [21, 40, 42, 44, 46, 47], especially as staff adjust to new ways of working; (2) the necessity of supporting staff by delivering advanced training on the neurobiology and behavioral sequelae of trauma and providing ongoing supervision, coaching, and debriefing of seclusions, restraints, and patient/staff injuries [20, 34, 40‐43, 45‐47, 49]; (3) the power of listening to patients and families about their experiences, needs, and priorities in the treatment process [21, 42, 47, 48]; (4) the importance of reviewing data and outcome indicators to motivate continued improvement [20, 40, 44‐46, 48]; and finally, (5) the need to align policy and practice, formal and informal, with the overarching principles of trauma informed practice [21, 40, 44‐46, 48]. After describing these five factors in greater detail below, we discuss our original implementation-science informed TIC program theory model and suggest changes to the theory based on this review.

Successful TIC implementation requires that organizational leadership, especially senior leaders, be visibly committed to the change process. This means that leaders change their own leadership practices to highlight organizational commitment and support for TIC [33, 47]. Across trauma informed care initiatives, staff knew TIC was a priority by the way leaders behaved. Senior leaders made TIC a standing item in high level meetings, allocated resources, set clear targets, communicated the rationale for the initiative with staff, and articulated “an unwavering belief” that TIC goals were achievable.

In their implementation of the Six Core Strategies, Caldwell et al. [42] underscored the importance of leaders in championing organizational change,

Rigid thinking and old-school mindsets of staff can result in minimal change. Leadership is key to addressing the rigid thinking and mindset of staff and should be outcome-focused to send the message to the organization that culture change is going to happen, the program is changing, and that staff can be part of this change or not (36).

Similarly, executives and leaders at Damar Services, a large residential treatment center, endorsed the agency’s shift to restraint elimination and modeled for staff that the shift in philosophy was not only “part of Damar’s new philosophy, but was the right thing to do as consistent with research and best practice for long-term outcomes” (5) [47]. Finally, two studies underscored the impact leaders can have on the success of TIC by conducting a thorough needs assessment and formulating a clear plan for implementation to guide the organization in achieving goals [21, 46].

While implementation science [50] stresses the importance of coaching over one-off training, most TIC frameworks and models in this review urged comprehensive staff training to help staff understand the purpose of TIC and to develop staff buy-in. Specifically, psychoeducation on the neurological and behavioural impacts of trauma was found to be critical [20, 41, 48]. The Risking Connection model and the Sanctuary model deliver curricula via a comprehensive staff-training module. Post-training measures demonstrated changes in staff knowledge, beliefs, and behaviour, although particulars were not reported. Furthermore, studies indicated that training is important because it gives staff common language to use regarding patient experiences and particular trauma informed interventions to be used with patients [33, 41].

Beyond training, studies included in this systematic review indicated the importance of staff members feeling supported throughout the change process. Recertification, ongoing training, coaching, and supervision reinforced trainings and provided staff support. For example, in a large residential facility in Indiana [47], a “resource team” was trained in behavior management and intervention techniques, with recertification required every 3 months. Additional trainings on best practices were provided for all employees of the facility, with direct-care staff required to be recertified in verbal de-escalation techniques every 6 months. In a study of the Attachment, Regulation and Competency (ARC) model [46], which produced a 50% reduction of restraint occurrences within the first 6 months of ARC implementation, researchers discovered that “Staff trauma responses impact staff and clients, as staff may be less able to effectively support and intervene with clients who are experienced as frightening or particularly difficult, as well as difficulty intervening all clients, because with of hypervigilance/hyperarousal” (683) [46]. Staff education included training in the Child–Adult Relationship Enhancement (CARE) model adapted from Parent Child Interaction Therapy (PCIT) to ensure staff trainings were “both didactic and experiential.” Trainings included hands-on opportunities for staff to practice self-regulation techniques and focused on the “developmental impact of trauma, building secure attachments, increasing self-regulation and competency, and self-care and vicarious trauma” (684) [46].

Most models included in this systematic review encourage the inclusion and participation of children and family members in care planning and treatment decisions. Although this element of successful TIC implementation seemed to occupy a less central role in the literature than we hypothesized (or was omitted from author discussion), consultation with patients and families was nonetheless discussed in depth by some authors.

For example, Caldwell et al. [42] reported that including youth and family was central to their success in preventing seclusion and restraint (see Table 3). This was, in part, because researchers and implementers invited youth to share their experiences of restraint with staff. Youth reported that restraint resulted in a loss of self-respect and dignity and in feeling less safe when watching peers. Holstead and colleagues [47] also involved patients in staff training so that staff could hear patients’ experiences of being restrained. In the development of their Trauma Informed Training Curriculum, Hummer et al. [21] emphasized child and youth choice and control, power sharing, collaboration, and caregiver involvement. Lebel et al. [33] suggested involving children and youth in debriefing critical incidents. Finally, the ARC model teaches family members psychoeducational, relational, and regulation techniques so that they can continue to use these skills when the child or youth is eventually discharged from the facility [46].

Across TIC implementations, an outcome orientation was promoted through regular data sharing in grand rounds and staff meetings. Across implementations, data comprised seclusion and restraint incidents, staff and patient injury rates, and diagnostic and functional symptom prevalence and severity. Data sharing was particularly germane to seclusion/restraint reduction initiatives. For example, Azeem et al. [40] report that outcomes were achieved and maintained by establishing seclusion and restraint reduction targets and goals, collecting and sharing real time data with units so they could monitor progress, encouraging friendly competition between units, and rewarding superior performance—both individually, via performance reviews, and collectively, by unit reviews. In complex initiatives, clinical improvements were also shared with staff to motivate them [48].

Across studies, consistent multilevel effort was required to align the milieu and organizational culture with the explicit principles of the chosen TIC model or philosophy. One way to bring about change of sufficient magnitude is to adopt a “therapeutic community” approach, such as the one promoted by the Sanctuary Model. In the therapeutic community model, the environment and culture of the organization are therapeutic tools themselves [48]. For example, organizations implementing TIC are encouraged to change the physical environment of the unit to make the treatment space feel safe and welcoming for both patients and staff [33]. Reviewed studies also suggested that trauma-informed principles be included in mission and vision statements, and that such statements be posted visibly to serve as reminders of TIC goals [21, 33].

With regard to changing organizational culture, Goetz & Trujillo [44] found that common challenges to successful implementation of their Patient Focused Intervention Model included troubleshooting staff opposition to longer times required to manage episodes of aggression, for example, through a “show of support” vs. a “show of force.” Eventually, “going hands on” came to be viewed as a de-escalation failure, indicating significant change in the culture of the unit. Additionally, Greene et al. [45] summed up the process required to align their model, Collaborative Problem Solving, with unit policies and practices:

The staff examined many long-standing unit policies and procedures, such as expectations for patient participation in therapy groups, visitation hours and policies, the grouping of patients, and staffing patterns, and worked together to improve compatibility between the unit structure, the primary goals of stabilization and assessment, the staff, and patients (612).

Findings from this review suggest that allocating process time for the slow and organic changes that must take place to accommodate the new way of practicing should be factored into TIC implementation plans.

Limitations of this systematic review included: (1) a truncated five-step realist review process in which we were unable to contact authors of all studies chosen for inclusion in the review to gather additional information about implementation context, mechanisms, and outcomes; (2) little description of our efforts to engage in knowledge translation with key stakeholders involved in the project of which this review was an initial component; and (3) no quantitative threshold for program/intervention quality/assessment of bias. Findings of the review should thus be approached with scepticism and applied with caution.

We found two large discrepancies between our original program theory model and the data we analyzed systematically: (1) the sequence of implementation activities undertaken, particularly activities to ensure patient and family participation and 2) the importance of choosing a particular program model (see Figs. 2, 3).

First, in our original program theory, community inclusion was viewed as a precursor to leadership commitment. That is, we speculated that successful trauma informed care would require that leaders of residential treatment centers, mental health agencies, health systems, and hospitals would first consult patients and families before deciding what changes to make or what model to adopt. However, overall, only a few authors [21, 42, 47, 48] discussed consultations with patients and families. Notably, when community inclusion was discussed, it had a very positive impact on the initiative—especially when patients spoke directly to staff about their lived experience.

Second, in our original program theory, we asserted that choosing a comprehensive evidence informed practice model would aid successful TIC implementation, especially when implemented with fidelity. But in fact, although some authors emphasized the salience of their particular model and credited it with changes which were achieved in staff and milieu behavior [20, 21, 41, 48] almost half the studies we reviewed reported efforts solely intended to reduce seclusion and restraint [34, 40, 42‐44, 49]. What’s more, despite a critique of standalone S/R reduction initiatives, these approaches produced significant reductions in episodes of seclusion and restraint, as well as staff and patient injuries (See Table 3). Following are brief profiles of two initiatives that represent these poles of trauma informed care: comprehensive TIC models and primary S/R reduction models.

In their discussion of Collaborative Problem Solving (CPS), a comprehensive model, Greene et al. [45] assert that reducing restrictive measures is not enough. Instead, a theoretically-based model should be adopted, as theory builds analytic capacity and increases staff understanding of the difficult behaviors they will encounter when working with children and youth:

Although reducing the use of timeout, quiet room time, restraint, and seclusion is an important goal, focusing on that specific goal alone is unlikely to accomplish the mission. Rather, we have found that even with a strong commitment from unit leadership to reduce or eliminate such practices, staff must also be provided with a comprehensive model of care, including a common set of assumptions about the factors underlying children’s aggressive or unsafe behavior, an understanding that the manner in which limits are set and expectations pursued by adults may precipitate such behavior, and an emphasis on crisis prevention rather than crisis management. In this view, reduction in the use of physically restrictive procedures is an outgrowth of good care, not necessarily an endpoint in and of itself (611).

Outcomes for this CPS implementation were very positive: Seclusion and restraint episodes declined from 281 in the 9 months before training to one incident 15 months post-training. Additionally, staff and patient injuries declined from an average of 10.8 per month to 3.3 per month.

On the other end of the spectrum, a modest UK intervention, Post Restraint Reduction Meetings, was quite straightforward [43]. Reduction Restraint workshops were delivered to all staff for children’s homes and residential settings. Additionally, reduction restraint meetings were held within 72 h of each restraint, and coaching from researchers was made available to staff. This intervention, exclusively focused on restraint reduction, achieved a 32% decrease in restrictive physical interventions—with the greatest reduction observed in the most restrictive supine floor restraints. These findings are consistent with those of Martin et al. [34], who suggest that “restraint may be a more achievable first target of reduction efforts” than seclusion or other targets for improvement (1409).

In the course of systematically reviewing studies, a pattern emerged in which the larger scale organizational cultural changes attempted by comprehensive models—which require more resources on the front end—perhaps produced longer-term and ‘deeper’ changes to organizational culture. This finding is consistent with implementation science literature, as ‘deeper’ organization change requires repeated and direct confrontation with “adaptive challenges” versus “technical problems” [51].

As discussed below, TIC initiatives may benefit from allocating dedicated staff time so that those implementing TIC are not recruited into “old” ways of doing practice or torn between roles because of time and resource allocation or role conflict. For example, discussing implementation of the Fairy Tale model in a particular agency, Greenwald et al. [20] noted,

It is the first author of this study’s impression that this agency’s therapists adopted the Fairy Tale model more slowly and incompletely than any other training cohort in recent years. This…seemed to be a direct consequence of having competing roles. When we have trained other therapists with similar dual roles, they had similar difficulty. The therapists’ time and role definition must be protected so that they are able to provide the treatment (150).

Articles reviewed suggested that to achieve a trauma informed organizational milieu which embodies patient choice, collaboration and control, organizations may confront long-standing issues like power struggles, the culture of psychiatry, and perceived efficiencies of using physical and chemical seclusion and restraint versus interventions that require substantial time and skill (e.g., collaborative problem solving). Brown et al. [41] found that staff who were trained in the Risking Connection model and who trained other staff showed sustained positive changes in knowledge, beliefs, and behaviors. Although not definitive, a train-the-trainer rather than purveyor model may produce financial efficiencies and generate longer term change in organizational culture.