Exploring factors that influence success when introducing "The Safewards Model" to an acute adolescent ward: A qualitative study of staff perceptions.

PROBLEM: The Safewards' model identifies factors that can lead to conflict and addresses these factors, using ten interventions, within inpatient mental health wards aiming to reduce "conflict and containment." The Department of Health (2014) and Care Quality Commission (2017) supported the use of Safewards to reduce restrictive practice across all mental health settings in the UK, but its application to adolescent mental health remains relatively unexplored. This study therefore aims to address the research question: "What are the factors influencing the success of ten Safewards' interventions when implemented onto an acute adolescent ward?"
METHODS: Eight healthcare assistants and two nurses who had attended Safewards' training participated in semi-structured interviews four months after Safewards was introduced to an acute adolescent ward. The interviews were transcribed verbatim and analyzed using thematic analysis. Data analysis was conducted inductively by developing data-driven themes.
FINDINGS: Many of the factors influencing Safewards' success in adolescent mental health (e.g., acuity; dependence on nonregular staffing; lack of leadership and operating procedures) paralleled the evidence found in adult services.
CONCLUSIONS: This study contributes new information by implementing "mutual help" and "calm down" principles with adolescents, as well as discussing barriers of operational procedures and benefits of patient involvement.

INTRODUCTION

Widespread use of restrictive practices within healthcare settings continues to come under scrutiny as they carry significant risk (Coles et al., 2015; MIND, 2013) and have a negative impact on recovery (Barnicot et al., 2017; Sullivan et al., 2013). Bowers (2014, p. 355) referred to the restrictive practice as “containment,” which covers “those (actions) staff do to prevent these (high risk) events from occurring” including: restraint, high level of observations, intramuscular injection, or making changes to the physical environment. The issue with restrictive interventions, such as restraint and intramuscular injection, is that they carry a certain level of risk (Coles et al., 2015; Innes & Curtis, 2015), whilst other measures, such as high‐frequency observations, are intrusive and infringe patients’ privacy and dignity (Pereira & Woollaston, 2007; Rooney, 2009). Reducing the need for restrictive practices, by preventing conflict, has been the focus of Bowers’ research for many years, leading to a large‐scale review of the literature (Bowers, Alexander, et al., 2014), informing the Safewards’ model (Bowers, 2014).

In relation to Safewards, Bowers (2014, p. 355) described “conflict” as “events that can threaten staff or patient safety,” which includes factors such as: aggression, suicide, self‐harm, substance misuse, absconding, and nonengagement in care. The Safewards’ model (Bowers, 2014) identifies the origins of conflict (see Figure 1), which Bowers (2014) used to inform 10 distinct interventions (see Table 1). The model identifies domains of: patient community; patient characteristics; regulatory framework; staff team; physical environment; and outside hospital. Bowers (2014) identified “modifiers” within these domains, highlighting ways patient and staff interactions trigger “flashpoints,” leading to conflict and thereafter use of containment. Interventions are aimed at reducing “flashpoints” by addressing the six conflict domains and associated staff or patient modifiers. A full description of the Safewards’ model and its ten interventions can be accessed through the Safewards’ website (Bowers, 2013).

Figure 1

Model provided by Safewards website: Bowers (2013)

Table 1

The Safewards interventions

1. Clear Mutual Expectations: Some challenging behaviors exhibited by patients are due, in part, to a lack of clarity around how they are expected to behave and a lack of consistency between staff about what those expectations are. Setting clear mutual expectations for both patients and staff allows patients to have control over what the expectations are and to understand the reasoning behind those expectations. This intervention was met on the adolescent ward by holding a meeting with young people to set agreed expectations. These were typed up onto a poster which was displayed on the ward.

2. Mutual Help meeting: Ward‐based conflict occurs from patient‐patient or patient‐staff interactions. A mutual help meeting promotes a supportive social community by promoting ways young people could support each other or ask for support from staff. This was addressed within the adolescent ward by hosting two meetings per day. “Morning meeting” and “sundown” (evening meeting) were introduced at the beginning and end of the day. The purpose was to allow staff and young people to set goals for the day and ask for help in achieving these goals. A mutual help poster was also designed by the young people detailing structured ways they can offer support to each other.

3. Positive words: Many handovers focus on patient behavior that is challenging or that which carries risk. With these handovers being time limited, less focus tends to be on positive aspects around patients care. To promote more balanced handovers, this intervention suggests something positive is said about each patient at handover. Positive quotes and a handover poster was displayed on the ward and nursing office, guiding nurses towards positive aspects of a patient's care. A poster was displayed above the phone with a flowchart indicating positive ways to handover to families following an incident. Staff training focused on positive handovers.

4. Soft words: Soft words are about changing our language to convey a softer message. Soft words poster was displayed in the ward office reminding staff of different ways we can approach challenging situations. Soft word message of the day was placed in the handover book to be read out at the end of handover.

5. Talk down: Talk down offers clear evidence‐based guidance on how staff can successfully deescalate an emotionally dysregulated patient. Talk down tips poster was displayed in nursing office and Safewards training offered guided talk down to support staff with successful de‐escalation.

6. Calm down methods: Calm down methods focus on ensuring positive coping skills are readily available for patients should they become distressed. Calm down and relaxation box were introduced to the ward allowing young people to use the box for relaxation and to facilitate grounding and sensory coping strategies.

7. Bad news mitigation: This intervention offers guidance on the best evidence‐based ways of delivering bad news to individuals. Bad news mitigation poster was designed and displayed in the nursing office.

8. Know each other: Know each other aids patients to build trusting relationships with staff and each other by offering information around everyone's hobbies and interests. A getting to know you book, detailing agreed information about the staff team, was designed and displayed on the ward for young people to read and get to know the people looking after them. The staff also gave patients the opportunity to write their own.

9. Reassurance: Many incidents on the ward cause heightened anxiety across the unit. Reassurance is about ensuring support is offered when needed and incidents are discussed with young people after, ensuring they are reflecting and getting appropriate support. To address this intervention with adolescents all young people were offered a debrief and a one‐to‐one talk with their nurse each day at morning meeting and sundown.

10. Discharge messages: A main patient characteristic that drove conflict was hopelessness. Practitioners and patients created a discharge tree and asked patients to write inspirational messages about their recovery journey on a leaf, placing this onto the tree. This was to help other patients to see that others have experienced positive outcomes.

Descriptions adapted from Safewards website: Bowers (2013).

Since the development of Safewards, many researchers have used the model, publishing research evidencing its effectiveness within adult acute, older persons, forensic, and psychiatric intensive care settings (Bowers et al., 2015; Fletcher et al., 2017; Stensgaard et al., 2018). In a randomized controlled trial, Bowers et al. (2015) used to conflict and containment checklists before and after Safewards’ implementation, finding a significant reduction in conflict events (15%) and use of containment (26.4%). Despite appearing to support the application of Safewards, the study drew criticism (Mustafa, 2015) due to the challenges it faced creating a truly “blind” condition, and the significantly low fidelity rates across experimental conditions (38%). Findings from qualitative research within adult healthcare have also identified difficulties with fidelity with staff identifying lower levels of perceived success when faced with: high acuity, staff shortages, lack of training, and poor leadership behind the approach (Fletcher, Hamilton, et al., 2019; Higgins et al., 2018; Price et al., 2016). Using a matched pairs design, Fletcher et al. (2017) were successful in measuring higher rates of fidelity, as well as demonstrating success with reduced seclusion rates (36%) in units using Safewards when compared to control wards matched for service and type. It is very difficult to determine whether the high rates of fidelity in this study were a result of successful implementation, or a methodological flaw whereby the fidelity checklists were more a measurement of evidence that was on display and not necessarily the degree to which the staff engaged with the display material (Bowers et al., 2015).

Using Safewards to reduce restrictive practice has also been the focus of reports by both the Care Quality Commission (2017) and the Department of Health (2014). The Care Quality Commission (2017) reviewed the efforts of five national health trusts to reduce the restrictive practice, encouraging services to consider evidence‐based approaches such as Safewards to support staff to change practice. The Department of Health (2014) outlined a guidance framework that aims to reduce restrictive practice on a larger scale, suggesting Safewards could be replicated across broader service settings whilst recommending for all providers to consider the implications of the Safewards’ model for their context.

Studies supporting Safewards’ application continue to grow within the literature, yet the evidence base supporting its use in adolescent services remains limited. Searching the databases (Delphis, Scholar & Cochrane) determined a relative dearth of research focusing on Safewards’ use within acute adolescent settings. Generalizing findings from such studies has also been difficult; research by Fletcher et al. (2017) examined Safewards across 13 wards, but only three of these were solely adolescent wards. The follow‐up studies used surveys to gain information based on staff (Fletcher, Hamilton, et al., 2019) and patient (Fletcher, Buchanan‐Hagan, et al., 2019) perspectives. Limited information could be derived around Safewards’ use with adolescents as the sample size was underrepresented with minimal staff (5.3%) and patient (6%) responses compared with adult services. Hottinen et al. (2020) were able to examine Safewards across six adolescent wards, but the focus of this study was on examining changes to the social climate, leaving a gap in the literature, with little being known about implementation barriers in relation to adolescent wards.

Relying on research derived from adult healthcare to guide practice with adolescents remains open to criticism due to fundamental differences in personal demographics, impacting both conflict and containment (Allan et al., 2017; Duke et al., 2014; Hert et al., 2011; Muir‐Cochrane et al., 2014). Despite these differences and regardless of gaps in the literature, reports from the Care Quality Commission (2017) and the Department of Health (2014) still support the use of Safewards across all settings as a strategy to reduce the restrictive practice. This study seeks to address gaps within the literature, by conducting semi‐structured interviews, following the implementation of ten Safewards’ interventions onto an acute adolescent ward.

Research question

What are the factors influencing the success of ten Safewards’ interventions when implemented onto an acute adolescent ward?

METHODS

Design

A semi‐structured interview, postintervention, design was used to gain qualitative data, eliciting staff views surrounding the introduction of Safewards four months after all ten Safewards’ interventions had been fully implemented onto the ward. This design accorded with the research question, allowing detailed information to be derived from the interview transcripts exploring factors influencing the success and also the challenges of the Safewards’ interventions.

Setting

The research was set in an acute adolescent inpatient ward where the staff had either no knowledge or only a basic understanding of Safewards and no experience of previous attempts to implement the strategy onto a ward. The unit was chosen as the principal investigator (NY) was employed on the ward as a registered nurse, a position that had been held for 14 months before initiating the project. Therefore, there had been a prior relationship with the participants, working alongside the team, and also had an involvement in aspects of general hospital training.

Safewards’ implementation

A twenty‐week staggered action plan was initiated, introducing one new Safewards’ intervention fortnightly, until all 10 interventions were successfully implemented. A longer approach was decided as, within the research, a 12‐week implementation plan (Fletcher et al., 2017) yielded higher fidelity scores then the previous 8‐week plan (Bowers et al., 2015). The decision to extend the implementation to 20 weeks was made through clinical judgment around what was practical for the ward, when taking into account staffing and resources. Each phase was initiated in a staff meeting, whereby an intervention was discussed, adapted, and, afterward, introduced to the patients. An email was circulated briefing staff about the interventions’ aims, objectives, and their role in ensuring tasks were carried out. A 4‐h Safewards’ training session was then delivered which taught staff the background to the model, supporting evidence, and its interventions. This study opted for a shorter training regime, as this was a more practical approach, whilst considering hospital resources, training procedures, staffing levels, and ward acuity. Since the principal investigator was working on the ward, they took full responsibility to implement all interventions with the support of a healthcare assistant. This approach differed from that of previous research (Bowers et al., 2015; Fletcher et al., 2017) and was chosen as the principal investigator had more time to commit to intervention delivery than the ward staff (due to acuity and staffing levels) and therefore felt that this method would achieve higher success. It was also an opportunity to research a slightly different implementation approach.

Participants and recruitment

Two nurses and eight healthcare assistants participated in interviews four months after the implementation of all the Safewards’ interventions. Participants were recruited at a ward meeting, where volunteers were asked to take part in semi‐structured interviews. Inclusion and exclusion criteria were set for the study; thus those who volunteered were only selected if they were nurses or healthcare assistants who were permanent staff employed before Safewards’ introduction and who had attended the training session. The sampling method was, therefore, purposive, as the researcher approached participants asking for volunteers, and participation was chosen based on set criteria. All participants that volunteered met the criteria so were included; there were no requests to withdraw data after the interviews were completed.

Data collection

The principal investigator took responsibility for conducting semi‐structured interviews and data analysis, following the delivery of training sessions and full Safewards’ implementation, the latter being assisted by a healthcare assistant. The principal investigator had previous knowledge and experience of Safewards through attending training sessions, experience on other units, and through academic‐based research. The framework for the interview (see Table 2), containing nine questions, was developed to meet the research question, by following the core aspects of Normalization Process Theory (NPT): establishing coherence, cognitive participation, collective action, and reflexive monitoring. The interview framework offered prompts, allowing the interviewee to elaborate on information from the previous answer. NPT was chosen to guide the interview as it is supported in the literature (May et al., 2009) as a useful tool to aid research into the application of intervention; it is, therefore, a good fit for the objectives of this study.

Table 2

Semi‐structured interview schedule

Name of interviewee ______________ Name of interviewer_____________

Participant number ______________ Job role: ______________________

During the initial Safewards’ introductory phase, how did the team adapt the principles to fit the wards current practice?

Examples: restructuring the ward timetable, changing displays, merging principles into the way the team practices.

Prompt: How did this influence the team's ability to consistently carry out the interventions?

As a nursing team do you think everyone understands the aims, objectives and expected benefits of the intervention?

Prompt: Did this influence the team's ability to consistently implement the interventions each day?

What are your responsibilities in ensuring the interventions are maintained?

Prompt: How does this impact the other responsibilities you undertake as part of your job?

Think about the Safewards’ interventions the ward has in place currently.

If staff did not implement these interventions throughout the day (e.g. not facilitate a morning meeting) what would the impact be on:

The staff

The patients

Which interventions do you believe had some success in achieving its aims to reduce conflict or restrictive practices?

Prompt: Were these maintained consistently after being implemented on to the ward?

Individual interviews were conducted face‐to‐face, preventing the influence of external factors on participant responses. Interviews took place off the ward, in an office, out of audible range, with only the principal investigator as the interviewer and the participant present. The duration ranged between eight minutes 23 s and just over 29 min. Interviews were audio recorded using a dictaphone and uploaded onto the computer as a secure audio file. Audio files were later transcribed verbatim with an emphasis on words, non‐verbal sounds, and additional input from field notes, adding contextual clarity (see Table 3). Non‐verbal data from field notes were not considered necessary to address the objectives of this study so were not included in data collection. Due to timing and scope of the project, repeat interviews, participant feedback from the transcripts, and sampling using data saturation, were not involved since they were not required to meet the aims of the research.

Table 3

Excerpts from transcript of P5 and P6 interviews

Data analysis

Braun and Clarke's (2006) approach to thematic analysis was used, guiding the researchers through a six‐phase process including familiarization with the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the report. This approach was followed as it provides a clear demarcated process by which to conduct a thematic analysis whilst ensuring that the flexibility of this type of methodology is maintained. Data analysis was, therefore, inductive; the initial aims of the project were set, and themes were developed and discussed. This allowed the identification of surface meanings within the data to address these meanings in relation to the research question. Quotes were then drawn from interviews to exemplify the themes. The principal investigator was responsible for data analysis, although each stage of the analysis was discussed with and reviewed by a coinvestigator (JL) who helped to collaboratively agree coding, and develop themes for discussion.

Ethical considerations

The study received ethical approval via the University of Southampton Ethics Committee through Ethics and Research Governance Online (Ethics ID number: 49640). Confidentiality, right to withdrawal, and study objectives were outlined using participant information sheets before arranging interviews, and consent forms were signed. The University of Southampton Data Protection Policy (2008) was followed throughout data handling, storage, and retention.

FINDINGS

The findings yielded four broad themes (see Figure 2): (1) Working with each other; (2) improved access to support; (3) the staff team; and (4) fitting Safewards into pre‐existing hospital processes. A number of subthemes emerged within each broad theme.

Figure 2

The coding framework [Color figure can be viewed at wileyonlinelibrary.com]

Working with each other

Before the implementation of Safewards several participants highlighted a disconnect between staff and patients describing there to be “a little bit of a war going on” (P5) with everyone seeming to be “on a different page” (P3). By introducing Safewards staff felt that this created a more cohesive ward community by allowing staff and patients to “work all together” (P5).

Staff also identified that by involving the patients in Safewards they felt “part of the puzzle” (P7), which helped to maintain Safewards, as patients would remind staff to lead on the interventions. Participants identified that both staff and patients contributed to the intervention “clear mutual expectations” through a community meeting where everyone gave idea's which helped to establish a set of expectations that was agreed upon by all. Participants felt that involving the patients made it easier for this intervention to be a success.

Adapting daily meetings by incorporating “mutual help” and “reassurance” principles facilitated healthy communication and ensured the patients were offered regular support. Participants discussed how these meetings gave them a “broader understanding of how the kids had done during the day” (P6), allowing staff to “offer support a lot easier, rather than expecting [patients] to come to [staff]” (P6).

Improved access to support

By introducing Safewards, patients gained access to support sooner as interventions such as “calm down” encouraged patients to use independent coping strategies. Before Safewards was introduced, participants felt the patients’ needs were not being met promptly as the staff was both “unaware of how the patients were feeling” (P5) and did not have access to the resources that “help them when they are feeling unsafe” (P5).

Support being offered before conflict occurs

Participants expressed that, before Safewards, “the team was reacting to the incidents as they happened, rather than before the young person went into crisis” (P8). This demonstrated that, as staff‐patient communication improved, staff were able to proactively resolve conflict by offering patients one‐to‐one support, or encouraging “calm down” skills, thus preventing a crisis from occurring. Staff identified that many patients on the ward often had difficulty approaching staff and asking for support. The introduction of the “mutual help” meetings promoted access to support by establishing forums where patients could communicate their feelings and encourage each other to talk with staff.

“[The mutual help meeting] … kind of helped us so we are prepared … having the morning meeting [helped the staff team to] know throughout the day [how patients were feeling] and it's not dragged out longer for them and we can kind of [support patients] quite quickly.” (P5)

Increase in access to independent ways of coping

Staff identified that, before Safewards, patients struggled to offer peers support without becoming emotionally dysregulated themselves. One staff member identified that patients were “supporting each other [in incident situations] that was triggering for them” (P4). Adapting “mutual help” by establishing clear boundaries around how patients can support each other, proved successful when working with adolescents as staff was able to offer effective ways they can support each other, without patients involving themselves with others, in times of crisis. Having access to “calm down” resources such as “sensory items” and “coping tools” (P9) was valued by staff as they allowed patients to independently self‐soothe which meant they were able to “de‐escalate themselves with just a little bit of prompting” (P1).

The staff team

High acuity, staff shortage, and the dependence on non‐regular staff left participants with an unmanageable workload, having to split limited time between ward responsibilities, managing conflict, and maintaining Safewards’ interventions. Many staff felt that once Safewards was introduced the job started feeling “full on at times” (P4), which led to there being “only a small number of people who were actively trying to implement [Safewards]” (P4). Staff also prioritized “all the roles and responsibilities of the job” (P4), which caused them to “feel a little at times like [Safewards] would become neglected” (P4).

Use of nonregular/agency staffing

Whilst critiquing Safewards, participants reflected on working alongside nonregular staff, identifying that when agency staff was used to fill leadership positions, they found it difficult to “teach them how to do Safewards in such a short space of time” (P8) as they ‘had not necessarily come in with an understanding of Safewards’ (P9) and therefore were “not leading on it” (P9). This created conflict in clinical decisions, with some staff maintaining fidelity towards Safewards and others opting for more restrictive practices, leaving patients feeling confused:

“It confuses the kids if I'm trying to … work according to an intervention and … my nurse in charge … has decided to do something different … I can't carry out the interventions and follow my nurse's orders … and quite often … I have to ignore the intervention”. (P6)

Contrasting views identified that interventions such as “clear mutual expectations,” and the way they were displayed on the ward, helped agency staff to practice consistently with the team.

“Having the … expectations up on the ward, where we do get a fair amount of agency [staff] … we can say to them, when they are being inducted … these are [the expectations] we've agreed on.” (P3)

Impact of short staffing and ward acuity on Safewards

Due to high acuity and staff shortages, often interventions such as “positive handovers,” “reassurance,” and “clear mutual expectations,” were viewed as less successful in achieving their aims. Not having the time to deliver a “positive handover” was discussed by staff with one participant identifying that after a busy shift the handover nurse “just wants to get the important [information] across to that they can go home” (P1). Limitations of time also impacted on patient care when offering support through “reassurance” in “mutual help” meetings; one staff member raised concerns that “there was more of a risk of not being able to have a one‐to‐one that is almost promised in that meeting” (P9). “Clear mutual expectations” set and agreed by patients and staff also “cannot always happen, because [staff] are busy” (P2), all of which led staff to highlight that without the time to commit to Safewards it becomes very challenging for the staff team to put “any of these practices in fully” (P1).

Fitting Safewards into pre‐existing hospital processes

To maintain staffing levels on the ward the hospital had a pre‐existing training structure, splitting training into multiple days, preventing the team from being trained together and limiting opportunities for staff to attend. This was discussed as a major implementation barrier with participants expressing a need for a team meeting to help maintain consistent implementation. Despite some hospital processes limiting intervention success, other participants discussed how, when interventions such as “mutual help” and “reassurance” were merged into the ward's structure (through morning and evening meeting), it reduced demand on staff time, leading to staff being more accepting of the interventions.

“I think they [the interventions] fit really well actually, I don't feel like its extra responsibility … I think it enables me to do a bit of a better job.” (P3)

DISCUSSION

When focusing on factors that influence the success of Safewards in adolescent services, this study found many similarities when compared to previous qualitative research undertaken in adult inpatient units (e.g., Fletcher, Hamilton, et al., 2019; Higgins et al., 2018; James et al., 2017; Price et al., 2016). James et al. (2017) identified that a busy chaotic ward, with staff shortages, unwell patients, and ward incidents were major barriers to the successful implementation of Safewards’ interventions. Low staffing also caused participants to perceive interventions such as “mutual help” and “calm down” as “unfeasible” in research by Price et al. (2016). The current study found similar results with “clear mutual expectations,” “positive handovers,” and “reassurance” being identified as less successful interventions in the presence of high acuity and low staffing, as staff did not have the time to commit to these interventions. Despite this, staff identified that merging principles of “mutual help” into the ward's current structure, through twice‐daily meetings, led to less demands on staff time, allowing this intervention to be more successful. Low staffing levels contributed to a dependence on nonregular and agency staff filling leadership or “nurse in charge” positions, who did not lead on Safewards’ interventions, which caused conflict with less senior staff attempting to maintain the principles. James et al. (2017) found similar issues with negative adaptations of certain interventions being made, due to a lack of understanding by temporary staff, resulting in a reduction in intervention quality. Staffing issues also impacted the interventions in research by Higgins et al. (2018), with high turnover and lack of senior leadership affecting continuity and leaving staff without guidance.

The theme “fitting Safewards into pre‐existing hospital processes” led to new insights, thus expanding on previous research. Operating procedures were perceived as barriers to successful implementation in several studies; these included rushed implementation processes, lack of ownership of the intervention, and a fragmented “train the trainer” process (Fletcher, Hamilton, et al., 2019; Higgins et al., 2018; Price et al., 2016). In the current study, participants focused on issues that arose from the team being trained separately, and not using intervention champions was also discussed. Bowers (2013) made specific recommendations that the team attend training together, which involves allocating “champions,” and splitting the workload, thus reducing demand on staff and facilitating discussion around how to merge the principles onto the ward. This approach has been used in previous research, documenting positive staff engagement and the successful implementation of interventions (Kipping et al., 2019; Hamilton et al., 2016). The use of “champions” showed limitations in one study when chosen staff were not perceived as senior enough to lead on interventions (Higgins et al., 2018). In relation to the current study, due to its scope and hospital training procedures, staff were trained separately. The principal investigator “championed” all interventions with support from one healthcare assistant, leading that member of staff to perceive the workload to be “full on at times,” and causing some interventions to be “neglected.” This led to multiple suggestions for a team meeting to facilitate discussion surrounding workload and engagement.

Information emerging from “working with each other” contributed an alternative perspective, differing significantly from previous findings. James et al. (2017) concluded that feedback from patients influenced intervention quality as staff accepted or abandoned interventions based on this feedback. The current study expanded the patient role by promoting active engagement during implementation, which helped to maintain interventions consistently each day.

A key finding with considerable differences when compared to previous research was patient attitudes toward “calm down” and “mutual help” principles; in one study, these interventions were considered less impactful as they were perceived as “patronizing” and “condescending” (Fletcher, Buchanan‐Hagan, et al., 2019). In contrast, adopting these approaches in the current study was considered highly valued as they “improved access to support.” The sensory items used as part of “calm down” were reported to be useful tools when offering support and provided patients with skills to manage independently. Mutual help meetings acted as a forum for patients to establish their feelings, improving group communication, which helped staff identify where support was needed. During participant interviews, it was evident that the interventions were beneficial as they encouraged independent ways of coping whilst creating an environment where patients felt able to appropriately support each other.

Conclusions

When compared to previous qualitative studies, it would be reasonable to conclude that many factors that influenced the success of Safewards within adolescent services paralleled research found in adult units (e.g., Fletcher, Hamilton, et al., 2019; Higgins et al., 2018; James et al., 2017; Price et al., 2016). Factors such as acuity, dependence on non‐regular staffing, lack of leadership, and operating procedures consistently produced problems, regardless of ward demographics. Despite many similarities being identified, this study also found key areas where participant views differed from previous research. In contrast to research by Fletcher, Buchanan‐Hagan, et al. (2019), the current study found principles of “mutual help” and “calm down” advantageous when implemented onto an adolescent ward as they “improved access to support,” facilitated communication, and helped consolidate independent ways of coping. It would therefore be appropriate to encourage practitioners to apply these principles when implementing Safewards onto an adolescent ward. Further recommendations can be made for practitioners to give patients an active role in Safewards as this increased consistency with implementation in the current study, as everybody was engaged in the approach.

From an organizational perspective, hospitals are encouraged to adapt or change procedures to maximize the effectiveness of Safewards. When comparing the barriers of success to that of a more successful training regime (such as in Kipping et al., 2019 and Hamilton et al., 2016), it is evident that training staff simultaneously is advantageous by ensuring the use of “champions,” sharing the workload, and facilitating discussion and engagement.

Strengths and limitations

Insider research was used to meet the objectives of the study due to its ability to maintain depth and credibility of data sources, whilst also ensuring that the research was practical and accessible to the researcher, which is not always the case with other approaches (Johnston et al., 2016). In this way, it was possible to gather in‐depth and rich information, specific to adolescents. Other studies have tended to collect data using surveys across a wider range of ward types, an approach that is more generalizable for different contexts (Fletcher, Buchanan‐Hagan, et al., 2019; Fletcher, Hamilton, et al., 2019). The current research was based on a relatively small sample of participants and in relation to Safewards’ implementation on just one ward. This makes the findings difficult to generalize to other settings as the factors impacting success may differ between adolescent services due to variations in staff teams, client groups, and other ward‐based pressures.

In addition, further critique by Greene (2014) has suggested that pre‐existing relationships can cause response bias whilst also identifying the effect that preconceptions of the setting and context have on researcher bias, creating a threat to objectivity. Within this study, not only did the principal investigator have pre‐existing relationships, but also took the lead in Safewards training and implementation, which could have led to further response bias throughout the participant interviews. Despite this potential issue, in the current study the pre‐existing relationships, combined with knowledge of the setting and context, were found to be more advantageous than problematic as they allowed participants to feel more at ease and helped to guide the interviews effectively.

In addition to implementation and interviewing, the principal investigator also took the lead on transcribing and coding the interviews manually, an approach which could be argued a further source of bias as the investigator could unconsciously look to validate an expected theme. To mitigate this the coding and themes were reviewed by a coinvestigator who was not involved in the implementation or interviewing. Data software was not used in this study as it was a relatively small‐scale study, and it was considered that detailed manual scrutiny of the transcripts both by the principal investigator and coinvestigator was more likely to yield insightful themes and a relevant coding schema. It could be argued that the use of software may have aided the research if the sample and number of interview transcripts had been larger, and a coding procedure such as line‐by‐line coding had been the approach of choice (Rademaker et al., 2012; Woods et al., 2016).

The study's methodology also has its advantages and disadvantages. Normalization Process Theory (NPT) was used to guide the interview format which was a good fit for the study and allowed the interview questions to be guided by robust clinical evidence (May et al., 2009). Coding the data into the four NPT constructs (coherence, reflex monitoring, cognitive participation, and collective action) was challenging. When organizing the data into themes it was evident that the codes were not necessarily demarcated in the same way. Some codes did not fit into one of the four categories and other codes overlapped into more than one. These issues have been highlighted in a systematic review by McEvoy et al., (2014) who recommended for researchers to be “critical” and “flexible” in their use of NPT in their work. For this reason, the research coding and data analysis adhered more closely to Braun and Clarke's (2006) approach to thematic analysis.

Although this study can make recommendations around factors influencing the success of Safewards when introduced to adolescent services, it cannot provide compelling evidence as to whether the strategy was effective at reducing conflict and containment. If the Department of Health (2014) and Care Quality Commission (2017) continue to recommend Safewards, then robust controlled trials are needed, specific to adolescent services, providing data as to whether overall restrictive practices are reduced.

CONFLICT OF INTERESTS

When carrying out this study the researcher was employed for the organization as a qualified nurse so held a professional working relationship with participants and employers. It is not considered that this relationship interfered with the ability to carry out the research without bias.