The effectiveness of brief non-pharmacological interventions in emergency departments and psychiatric inpatient units for people in crisis: A systematic review and narrative synthesis

Abstract

Objective:

Heterogeneous brief non-pharmacological interventions and guidelines exist to treat the burgeoning presentations to both emergency department and inpatient settings, for those in a crisis of mental ill-health. We systematically reviewed the literature to create a taxonomy of these brief non-pharmacological interventions, and review their evaluation methods and effectiveness.

Method:

We conducted a systematic review across Cochrane, CINAHL, DARE, Embase, MEDLINE, PsycINFO databases. Studies meeting quality criteria, using Joanna Briggs Institute tools, were eligible. Interventions were categorised, and outcomes synthesised.

Results:

Thirty-nine studies were included: 8 randomised controlled trials, 17 quasi-experimental, 11 qualitative studies, and 3 file audits. Taxonomy produced six coherent intervention types: Skills-focussed, Environment-focussed, Special Observation, Psychoeducation, Multicomponent Group and Multicomponent Individual. Despite this, a broad and inconsistent range of outcome measures reflected different outcome priorities and prevented systematic comparison of different types of intervention or meta-analysis. Few brief non-pharmacological interventions had consistent evidential support: sensory modulation rooms consistently improved distress in inpatient settings. Short admissions may reduce suicide attempts and readmission, if accompanied by psychotherapy. Suicide-specific interventions in emergency departments may improve depressive symptoms, but not suicide attempt rates. There was evidence that brief non-pharmacological interventions did not reduce incidence of self-harm on inpatient wards. We found no evidence for frequently used interventions such as no-suicide contracting, special observation or inpatient self-harm interventions.

Conclusion:

Categorising brief non-pharmacological interventions is feasible, but an evidence base for many is severely limited if not missing. Even when there is evidence, the inconsistency in outcomes often precludes clinicians from making inferences, although some interventions show promise.

Keywords

Emergency service hospital psychiatry crisis intervention

Introduction

Background

Psychiatric presentations to emergency departments (EDs) continue to rise (Hill et al., 2021; Holland et al., 2021; Sara et al., 2022), leading to challenges related to ‘boarding’ in the ED that have proven difficult to address (Nordstrom et al., 2019; Zun, 2012). Many EDs rely heavily on ED staff to manage these presentations; however, various alternative system approaches have emerged (Coates, 2018; Fitz-Gerald and Takeshita, 2020). In Australia, one common model involves short-stay emergency psychiatric wards situated either adjacent to or within the ED. These wards are staffed with specialised psychiatric teams responsible for serving both the ED and the short-stay ward, known in Australia as Psychiatric Emergency Care Centres (Frank et al., 2005; Seymour et al., 2020) or Psychiatric Assessment and Planning Units (Mitchell et al., 2020). Similar services, known as Psychiatric Emergency Services (Haack et al., 2020), have been established in the United States, and the United Kingdom is also starting to adopt them (Sampson et al., 2022). Research supports their effectiveness in reducing the length of stay (LOS) in the ED (Anderson et al., 2022). Importantly, these units differ from outpatient or ‘partial programme’ settings, as they focus on brief admissions before either discharging to the community or transferring to longer-stay wards. These service models also offer the opportunity to ensure ethical, efficient and effective care by initiating psychiatric treatment within the ED and possibly continuing it through to the short-stay inpatient space.

Of course, effective care is vital not only to ensure efficiency. Central to our understanding are the people who are reaching out for help. People presenting to EDs may experience interpersonal, environmental, and system-related factors (Navas et al., 2022) that impact their willingness to return to EDs (Rosebrock et al., 2022). If we are to provide excellent person-centred care, we must understand how the ED and inpatient spaces can be used in an evidence-based way.

The specific brief interventions used in this population and setting can be broadly thought of as pharmacological (e.g. sedation) and non-pharmacological interventions (NPI). There are systematic reviews that evaluate brief pharmacological treatment for mental health crises in the ED showing, e.g., that ketamine, droperidol and midazolam are the safest options for managing severe agitation (deSouza et al., 2022). There are systematic reviews of longer-term NPIs started in the ED that are then continued in the outpatient setting, despite the very great differences between environments (Chaudhary et al., 2020; Johnston et al., 2019; McCabe et al., 2018; Ougrin et al., 2012; Wilson et al., 2020) which indicate that NPIs such as green or gold card clinics and postcards to people after self-harm attempts may show potential in reducing suicide and suicide attempts. Other reviews have evaluated environmental interventions in the ED (MohammadiGorji et al., 2021) showing that environment plays a central role in mitigating aggression, service-wide interventions (Baker et al., 2021) showing great need for better systems to manage restrictive practices, psychological therapies on adult inpatient units (Jacobsen et al., 2018) showing that inpatient research lacks rigour and organisation, and the evidence for environment, policy and practice changes addressed acute severe behavioural disturbance in the ED (Weiland et al., 2017) demonstrating a deficit in research surrounding acute severe behavioural disturbance.

However, a critical gap remains in systematic advice concerning brief non-pharmacological interventions (bNPIs) that could be offered by psychiatric teams across both the ED/short-stay inpatient environments, meaning that the goals and effectiveness of bNPIs that are delivered within a very brief timeframe remain largely unexplored. This is of clinical significance because so many clinicians work across this interface (Coates, 2018; Evans et al., 2019), particularly in Australia. Furthermore, an appraisal of international guidelines for emergency mental health care in hospital settings written 21 years ago (Allen et al., 2002; National Institute for Health and Care Excellence (NICE), 2022) emphasised the need for such a review. Peak bodies urge ‘a written protocol that specifies the most effective and least restrictive approaches to common behavioural emergencies seen in the [emergency] service’(Allen et al., 2002: 23), and that a ‘psychological intervention that is specifically tailored for adults who self-harm’ should be provided in an environment where ‘healthcare staff should be appropriately trained . . . in the therapy they are offering’(NICE, 2022: 33). Without an understanding of the underpinning evidence, such guidelines cannot be followed consistently. Given over 20 years have elapsed since this first call for evidence, it is critical this is now addressed. With steadily increasing service demand, Australian ED and short-stay inpatient units are flying blind.

For this area of study to progress, a clear ‘map’ of crisis interventions that can be ‘specifically tailored’ in these environments is imperative. This will enable a comprehensive understanding of which interventions are needed for which person and at what time (Kleiman et al., 2022), and ensure the ED and ward environments are used as efficiently as possible (Kleiman and Nock, 2020). This may give rise to interventions that can be brought from the inpatient to the ED space, and vice versa. To date, a systematic review of this scope has not been conducted, possibly because (1) bNPIs are highly heterogeneous, making this an ambitious project, and (2) Emergency Psychiatry is not yet ‘organised’ as a treatment group in the same way as other areas of psychiatry. Nonetheless, as a starting point, we must have the evidence summarised, synthesised and organised so that a path forward may be established.

Objectives

To systematically review the evidence for brief NPIs in EDs and psychiatric inpatient units by:

Creating a taxonomy of these interventions;

Appraising and synthesising outcome measures used to evaluate them; and,

Systematically reviewing their effectiveness.

Methods

Protocol and registration

This review was registered with PROSPERO (CRD42021287496).

Criteria for considering studies for this review

Randomised controlled trials (RCTs), non-RCTs, cohort and case–control studies, case series and case reports, surveys and qualitative studies were included. We planned to include non-randomised studies because of the expected dearth in RCTs (Johnston et al., 2019), even though this increases the risk of bias. We included qualitative studies because these help guide and interpret quantitative data. See Supplementary S1 for PRISMA flowchart and checklist. Risk of bias was measured using Joanna Briggs Institute Risk of Bias Checklists, with a plan to exclude high-risk studies.

Selection criteria

Population

Study participants were persons who were in crisis of mental ill-health. There is no standard definition of in crisis in the literature, thus we defined participants as being in crisis when they were presenting to ED with any complaint related to mental or behavioural health, or an inpatient on a psychiatric ward having self-harm thoughts/behaviours or agitation/aggression. This definition reflects the nature of crisis as time-limited and where an important change has occurred. Studies of interventions for people solely with drug- and alcohol-related presentations were excluded.

Interventions

All primarily brief NPIs aimed at addressing psychiatric complaints were included. Incidental medication use was not an exclusion criterion, but studies were excluded if medications were identified as a component of the intervention being evaluated. Given the short duration of admissions to crisis units, we excluded interventions that were longer than 1 week. Interventions that were used during crisis admissions, even if they were not used on a crisis unit, were included. We excluded seclusion and restraint-only papers but a priori included papers (and stratified results accordingly) where other bNPIs were used. We included solely clinical interventions, rather than processes of care pathways.

Treatment setting

We included studies that described treatment within EDs or psychiatric wards, including those that started in the ED and continued in inpatient settings. We excluded interventions that were commenced in the ED and continued in outpatient settings, e.g., brief courses of therapy because reviews of these already exist (Chaudhary et al., 2020; Johnston et al., 2019; McCabe et al., 2018; Ougrin et al., 2012; Wilson et al., 2020). We excluded bNPIs conducted on non-psychiatric (medical) wards; although these patients may well be in crisis, we expected that the interventions would be different due to the lack of mental health trained staff. We excluded studies conducted in aged care facilities, group homes, jails and other non-hospital settings.

Outcome measures

We included papers with either quantitative or qualitative analysis. No outcome measures were excluded.

Articles

Articles were those written in English, in peer-reviewed journals, without date restriction. Searches were performed between 25 August 2021 and 8 December 2021, and again on 7 September 2022. Studies not yet complete were excluded.

Search methods

The database search was carried out by one researcher in Cochrane Central Register of Controlled Trials (CENTRAL), CINAHL, DARE, Embase, MEDLINE and PsycINFO, and peer reviewed by a librarian specialising in systematic reviews. Search criteria for each database are found in Supplementary S2. We excluded all evidence syntheses, expert opinion and descriptive studies; the reference lists of the former were hand-searched. Researchers also reviewed government health websites, including UK NHS, Canada, NZ and USA for references, plus key non-government organisation crisis resources for relevant articles.

Method of inclusion/exclusion

All studies were title screened by one author (J.P.H.). Four authors then reviewed abstracts of all screened inclusions, and conducted full text reviews, with each paper being reviewed by two authors independently for content and quality (J.P.H., M.C.B., L.M.N. and S.M.H.). When disagreement arose, authors discussed inclusion versus exclusion explicitly. This was particularly the case when differentiating between ‘service’ and ‘clinical’ interventions, as many interventions take into account the structure, staffing, models of care and care pathways. These studies were discussed at length and disputes adjudicated by A.M. and N.G.

Data extraction and management

One author (J.P.H.) systematically extracted data from all papers using Covidence and Microsoft Excel. Two authors (L.M.N. and M.C.B.) independently extracted the data and checked it against the original author’s data.

The Cochrane Handbook for Systematic Reviews of Interventions (Higgins et al., 2019) and the JBI Manual Evidence Synthesis Chapter 8: Mixed methods systematic reviews (Lizarondo et al., 2020) guided the process of data extraction and management. We piloted a data extraction form in Covidence using a pre-specified tabular format. Supplementary Table 2 shows itemised data for each paper after piloting multiple data collection formats. If data were missing, the authors were contacted in order of authorship. Data extraction was first conducted on 15 December 2021.

Intervention taxonomy

There was a high level of clinical heterogeneity of NPIs. To organise the interventions in a way that is helpful for clinicians, and potentially allow for data synthesis, the interventions were classified into a taxonomy. Multiple iterations of this taxonomy were discussed by three authors (J.P.H., N.G. and A.M.). The ‘Categories’ were created to provide a structure by which to understand the focus, form and content of the intervention, on the premise that similar interventions are expected to have similar theories of change. There was no a priori specific grouping of intervention prior to extracting the data. After being classified into Categories, similar interventions were then further grouped into intervention ‘Types’ (Supplementary Table 3 and Figure 1).

Figure 1.

Taxonomy of NPIs found in EDs and inpatient units for people in crisis. Measures of treatment effect.

Data were selected for reporting:

If there was a primary outcome defined in the paper, this was used;

If no primary outcome was reported, then any outcome used consistently across studies within each category;

If an effect size was reported in the paper, this was extracted;

If no effect size was reported, then effect size was calculated where possible.

Effect size types reported by papers assessing quantitative outcomes include partial η², Cohen’s d and odds ratios (ORs). Effect size types calculated by the authors of this paper include Cohen’s d and ORs. For continuous data, Cohen’s d was calculated by subtracting the means of the two groups and dividing by the pooled standard deviation (SD) (Lenhard and Lenhard, 2016). For groups of different sample sizes, the pooled SD was adjusted with weights for the sample sizes (Lenhard and Lenhard, 2016). For categorical data, ORs were calculated. The data used for calculation of effect sizes are presented in Supplementary Table 4. There were insufficient data to perform a sensitivity analysis.

The intervention types that were deemed to be similar were grouped and analysed together.

Results

Included studies

Thirty-nine studies met inclusion criteria. Eight studies were RCTs, 17 were quasi-experimental, 11 were qualitative, and 3 were file audits. Thirteen studies were high quality (Tables 1 –6) but no interventions had sufficient data to allow for meta-analysis. To aid understanding of upcoming sections, readers should repeatedly refer to Figure 1.

Table 1.

PICO table of ‘Skills-focussed’ intervention category. Studies are ordered by study type (randomised controlled trial, quasi-experimental) and date (most to least recent).

Author	Patient population and context	Intervention	No. of intervention episodes	Study type	No. of participants	Measure	Effect size or other reported outcomes	Risk of bias
‘Skills-focussed’ intervention category
Myhre et al. (2018)	Adults in first week of admission on inpatient ward	Behavioural activation: self-monitoring, activity scheduling, assessment of goals and values	5 per participant, over less than 1 week	Prospective randomised controlled trial for feasibility	ControlN = 9Intervention N = 10	Number of sessions completed	Mean = 4	Moderate
Myhre et al. (2018)	Adults in first week of admission on inpatient ward		5 per participant, over less than 1 week	Prospective randomised controlled trial for feasibility	ControlN = 9Intervention N = 10	Beck Depression Inventory II (Norwegian)	Partial η² effect of time = 0.88Partial η² effect between groups = 0.22	Moderate
Huffman et al. (2014)	Adults who are suicidal on an inpatient ward	One or more ‘positive psychology’ exercise:Gratitude letterPersonal strengths survey and writingActs of kindnessImportant, enjoyable, meaningful activityCounting blessingsWriting about ‘best possible self’ and how to achieve this	1 or more per participant	Quasi-experimental before/after study	Recruited: N = 61Completed at least one exercise: N = 52	Number of sessions completed	85% completed at least 1 exercise	Moderate
						5 point Likert scale of for each of 9 exercises for: Optimism	Cohen’s d ranges from 0.54 to 1.12 depending on the exercise
						Hopelessness	Cohen’s d ranges from 0.59 to 1.19 depending on the exercise

PICO: Population, Intervention, Control, Outcomes.

Table 2.

PICO table of ‘Psychoeducation’ intervention category. Studies are arranged according to study type (quantitative then qualitative) and then alphabetically according to author.

Author	Patient population and context	Intervention	No. of intervention episodes	Study type	No. of participants	Measure	Effect size or other reported outcomes	Risk of bias
‘Psychoeducation’ intervention category
Dyckman et al. (1999)	Adults presenting to ED with panic symptoms	Brochure: written psychoeducation about panicContact: usual care + brochure + ED consultation with educator specialising in panic	1 per participant	Quasi-experimental design. Retrospective control group constituted patients with panic from a medical records audit. 2 prospective intervention conditions assigned using quasi-random allocation.	Control: N = 269BrochureN = 53Contact: N = 32	Mean presentation rate to ED per year	Rate ratios for: Contact vs control = 0.675 (p = 0.07)Brochure vs control = 1.67 (p = 0.04)Contact vs brochure = 0.40 (p = 0.00)	Moderate
Gebhardt et al. (2022)	Veterans on a veterans inpatient ward	Group psychoeducation programme on understanding suicidal thoughts and behaviours, run by facilitators, using verbal and written cues.	1 per participant	Quasi-experimental design, with post-intervention evaluation.	N = 78 (phase 2)	5 point Likert scale of distress	Mean (SD)New information = 3.84 (1)Useful information = 3.99 (0.85)No change in distress = 49.3%Felt comfortable sharing = 44%	Moderate
Ward-Miller et al. (2019)	Adult females on an inpatient ward who had a history of self-injurious behaviours	Psychoeducation programme on teaching resourcefulness as alternative to self-injury, using hand puppets.	1 per participant	Quasi-experimental before/after study.	N = 16	Resourcefulness Skills Scale	Cohen’s d = 0.14 (95% CI: −0.84 to 1.12)	Low
Petrakis and Laxton (2017)	Adults admitted to an inpatient ward with First Episode Psychosis	Standardised psychoeducation session to help family and carers understand psychosis and psychiatric service	1 per participant family and carer group	Qualitative	N = 27		N/A	Low

ED: emergency departments; CI: confidence interval; PICO: Population, Intervention, Control, Outcome.

Table 3.

PICO table of ‘Environment-focussed’ intervention category.

Author	Patient population and context	Intervention description	No. of intervention episodes	Study type	No. of participants	Measure	Effect size or other reported outcomes	Risk of bias
Sensory modulation rooms
Lloyd et al. (2014)	Adults in distress on inpatient ward	SMR^a. Either patient or nurse initiated. Unclear if staff present.	N = 71 ‘client numbers’*	Quasi-experimental before/after, plus control for rate of seclusion	N = 71 ‘client numbers’	10 point Likert scale of distress reported by patient	Partial η² = 0.54	Moderate
Lloyd et al. (2014)	Adults in distress on inpatient ward		N = 71 ‘client numbers’*		N = 71 ‘client numbers’	Rate of seclusion	Partial η² = 0.24	Moderate
Novak et al. (2012)	Adults in distress on inpatient ward	SMR^a. Unclear if staff present.	75	Quasi-experimental before/after	Not recorded	10 point Likert scale of distress/anxiety reported by patient	Distress: Cohen’s d = 1.01Anxiety: Cohen’s d = −0.14	Moderate
Novak et al. (2012)	Adults in distress on inpatient ward	SMR^a. Unclear if staff present.	75	Quasi-experimental before/after	Not recorded	Number of seclusions per month	Cohen’s d = −0.14	Moderate
Seckman et al. (2017)	Adolescents in distress inpatient ward	SMR^a. 1:1 with ward staff.	Unclear	Quasi-experimental before/after	N = 89	11 point Likert scale of distress/anxiety reported by patient	Distress: Cohen’s d = 1.25Anxiety: Cohen’s d = 1.19	Moderate
Seckman et al. (2017)	Adolescents in distress inpatient ward	SMR^a. 1:1 with ward staff.	Unclear	Quasi-experimental before/after	N = 89	Number of seclusions per 1000 patient days before/after	Reduction by 32.8%	Moderate
West et al. (2017)	Adolescents in distress inpatient ward	SMR^a. 1:1 with multidisciplinary ward staff.	102	Quasi-experimental before/after	N = 56	11 point Likert scale of patient distress reported by patient/staff	Patients: Cohen’s d = 1.3Staff: Cohen’s d = 1.3	Moderate
West et al. (2017)	Adolescents in distress inpatient ward	SMR^a. 1:1 with multidisciplinary ward staff.	102	Quasi-experimental before/after	N = 56	Rate of seclusions	Cohen’s d = 0.2 (95% CI: −0.08 to 0.66)	Moderate
Sutton et al. (2013)	Inpatient adults who have used an SMR; clinicians who have helped use SMR	SMR^a. 1:1 with ward staff and alone.	Not recorded	Qualitative	Clinicians = 40Service users= 20	Thematic	N/A	Moderate
Sensory modulation
Dickson et al. (2021)	Adults presenting to the ED with psychiatric complaints	15 pound (6.8 kg) weighted blanket used for either 15 minutes or 30 minutes	1 per participant	Randomised controlled trial – control groups before/after	Control groupN = 515-minute intervention groupN = 430-minute intervention groupN = 6	PROMIS Anxiety and Anger short forms	No Cohen’s d were statistically significant. See Supplementary Table 4.	Moderate
Quiet rooms
Ray et al. (2017)	Adults who are agitated or violent on an inpatient ward	‘Psychiatric monitoring and intervention’: patient allowed to remain in their room alone with door closed. Staff respond to agitation.	Not recorded	Retrospective file audit	Total n = 426N people on PMI = 189	Change in rate of episodes of ‘continuous special observation’	Increase by 0.05 episodes per month prior to 2005; Reduction by 0.07 episodes per month after November 2005.	Moderate
Ray et al. (2017)	Adults who are agitated or violent on an inpatient ward		Not recorded	Retrospective file audit	Total n = 426N people on PMI = 189	Mean duration of episodes of ‘continuous special observation’	84% reduction in mean number of hours comparing 2002 with 2011.	Moderate
Glod C et al. (1994)	Children in QR because ‘threatening or assaultive behaviour or failure to respond to less restrictive interventions’, on inpatient ward	QR modified to have pastel pink walls, mural of sea animals, dark carpet.	30 instances Modified QR46 instances control QR, participants acting as own control group	Non-randomised control trial, combined 2 data collection periods, each 6 weeks long, 1 year apart	N = 10 (phase 1)N = 9 (phase 2)	Overt Aggression Scale of Yudofsky: Total Aggression score: mean time to reduce score by half	Total aggression scores were 45% lower in the modified quiet room than in the standard quiet room, F(l, 18 = 6.09, p < 0.03.	Moderate

‘Environment-focussed’ studies are ordered by intervention type (‘sensory modulation rooms’, ‘sensory modulation’ and ‘quiet rooms’), study type (randomised controlled trial, quasi-experimental, qualitative); patient population (adult, adolescent), date (most to least recent) and alphabetically by author last name, respectively.

SMR: sensory modulation rooms; CI: confidence interval; ED: emergency departments; QR: quiet rooms; PICO: Population, Intervention, Control, Outcomes.

SMR with multiple proprioceptive, tactile, visual, gustatory and olfactory items (list of items provided in paper). ‘Unclear if this indicates the number of intervention episodes or the number of people studied.

There were 71 clients, but it was unclear whether each client received one or more than one intervention.

Table 4.

PICO table of ‘Special Observation’ intervention category.

Author	Patient population and context	Intervention description from study	No. of intervention episodes	Study type	No. of participants	Measure	Effect size or other reported outcomes	Risk of bias
Stewart et al. (2009)	Adult inpatients who had/had not self-harmed, and had/had not received the intervention, on an inpatient ward	‘Keeping a patient within eyesight or reach’.	N/A	Retrospective file review	6328 patients were admitted over the period; 2144 patients included in the self-harm and/or SO portion of analysis.	Number of people on SONumber of people who self-harmed	Odds ratio likelihood of self-harm whilst on SO = 1.44	Low
Pitula and Cardell (1996)	Adult inpatients who had recently experienced SO as an intervention for ‘suicidality’, on an inpatient ward	Constant observation of patients by observing staff, including registered nurses, Mental Health ‘technicians’, and minimally trained lay workers or ‘sitters’.	N/A	Qualitative	14	Thematic	N/A	Moderate
Cleary et al. (1999)	Registered nurses on one of 4 adult inpatient psychiatric wards	A ‘registered nurse remaining continuously at arm’s length from, or within sight of (as specified), the patient; or a registered nurse observing the patient within a 30-minute interval’.	N/A	Qualitative	10	Thematic	N/A	Moderate

Studies are ordered according to study type (quantitative and then qualitative), and then participant (patients, then nurses).

SO: Special Observation; PICO: Population, Intervention, Control, Outcomes.

Table 5.

PICO table of ‘Multicomponent Group’ intervention category.

Author	Patient population and context	Intervention description	No. of intervention episodes	Study type	No. of participants	Measure	Effect size where calculable/available; primary outcome results where effect size not calculable from available data	Risk of bias
Silverman (2019a)	Adult inpatients	1: Group music therapyTherapist played song, discussed lyrics within the group, and Linked lyrics with recovery (ELA)2: Group music therapyGroup song-writing (ESW) with focus on recovery goalsControl: rock and roll bingo	1 per participant as part of a group, 45 minutes	Prospective randomised controlled trial, however, measures were applied before the control condition, and then after the 2 interventions. Thus, comparisons are not ‘intervention’ vs ‘control’, but rather ‘intervention’ vs ‘no intervention’.	ControlN = 23Intervention 1. N = 17Intervention 2. N = 29	Stages of Recovery Instrument	Cohen’s d between groups on ‘Moratorium’ scale:ELA vs no intervention = 0.56 (95% CI: −0.08 to 1.21)ESW vs no intervention = 0.19 (95% CI: −0.36 to 0.7)‘Awareness’ scale:ELA vs no intervention = 0.28 (95% CI: −0.361 to 0.911)ESW vs no intervention = 0.42 (95% CI: −0.14 to 0.99)‘Preparation’ scale:ELA vs no intervention = 0.47 (95% CI: −0.175 to 1.1)ESW vs no intervention = 0.28 (95% CI: −0.29 to 0.83)	Moderate
Silverman (2019b)	Adult inpatients	As in Silverman (2019a)	1 per participant as part of a group, 45 minutes	As in Silverman (2019a)	Control N = 29Intervention 1 N = 35Intervention 2 N = 28	Coping Self-Efficacy Scale reduced-item	‘Use problem-focussed coping’ subscale:Partial η² = 0.032‘Stop unpleasant emotions’ subscale:Partial η² = 0.009‘Get support from family and friends’ subscale:Partial η² = 0.017	Moderate
Barker and Dawson (1998)	Adult inpatients	Intervention: Group session with dog and dog handlerControl: Therapeutic recreation group session	1 per participant as part of a group, 30 minutes	Quasi-experimental pre–post design, wherein patients first attended the experimental intervention, and the next day attended the control intervention	Control N = 137Intervention N = 93	State Trait Anxiety Inventory	Cohen’s d within groups:InterventionMood disorders = −0.41 (95% CI: −0.81 to −0.01)Psychotic disorders = 0.41(95% CI: −0.93 to 0.11)ControlMood disorders = −0.33 (95% CI: −0.64 to −0.02)Psychotic disorders = -0.22 (95% CI: −0.65 to 0.22)Cohen’s d between groups:Mood disorders intervention vs control = −0.10 (95% CI: −0.54 to 0.35)Psychotic disorders intervention vs control = 0.05 (95% CI: −0.44 to 0.55)	Moderate
Walsh (1993)	Adolescent inpatients with suicidal ideation	Intervention: Suicidal adolescents formulate plans for future identity, created future self-image caricature posterControl: TAU	1 per participant as part of a group, 3 hours	Quasi-experimental non-randomised controlled trial.	Control N = 18Intervention N = 21	Beck Depression Inventory	Cohen’s d within groups:Intervention at posttest 1 Cohen’s d = −0.66 (95% CI: −1.58 to 0.27)Intervention at posttest 2 = −0.93 (95% CI: −1.88 to 0.016)Cohen’s d between groups:Intervention vs control at posttest 1 = 0.09Intervention vs control at posttest 2 = −0.14(calculated from data)	Moderate

Studies are ordered first by similarity of intervention (music therapy is the only intervention studied more than once), study type (randomised controlled trial, then quasi-experimental trial) and then by date.

CI: confidence interval; PICO: Population, Intervention, Control, Outcomes; ELA: Education Lyric Analysis; ESW: Educational Songwriting; TAU: Treatment As Usual.

Table 6.

PICO table of ‘Multicomponent Individual’ intervention category, ‘suicide-specific’ intervention type.

Author	Patient population and context	Intervention description	No. of intervention episodes	Study type	No. of participants	Measure	Risk of bias
No suicide contract intervention type
Bagatelas et al. (2022)	Adult inpatients expressing suicidality	Standardised, safety plan completed with ward staff, reviewed by psychiatrist or psychologist.	1	Quasi-experimental before/after	95	Number of people readmitted to ED	Moderate
Drew (2001)	Adult inpatients engaging in self-harm behaviour	‘Documentation of . . . agreement . . . between a health care provider and patient . . . patient agrees to talk with a staff member before acting on suicidal thoughts’.	1	Retrospective chart review	577 discharges	The likelihood of self-harm whilst on an NSC	Moderate
Potter et al. (2005)	Nursing staff on an inpatient ward caring for adults	Safety agreement tool – 4 questions to patient with multiple choice answers regarding thoughts of self-harm; strategies and instructions for nursing staff.	Once during each nursing shift for the first 72 hours	Quasi-experimental survey and qualitative (mixed-methods) study	41 nursesPatients sample size not reported	Mean per capita rate of self-harm per quarter (3 months) before vs after intervention was introduced to ward	Moderate
Specialised suicide-specific therapy intervention type
Dimeff et al. (2021)	Adults with suicidal ideation or behaviour in ED	Suicide specific interventions delivered by a chat bot via app.	1	Prospective randomised controlled trial	Intervention N = 14Control N = 17	Suicide-related coping scaleLikert satisfaction scale 1–5	Low
Ghahramanlou-Holloway et al. (2020)	Adults with suicidal ideation or behaviour on a military unit	Post-Admission Cognitive Therapy	5 × 60- to 90-minute sessions per participant over 3 days	Prospective randomised controlled trial	Intervention N = 12ControlN = 12	BDI	Moderate
Rotheram-Borus et al. (1996)	Adolescent suicide attempters and their mothers in the ED	20-minute videotape, with soap opera format plus structured family therapy session	1 per family group	Prospective quasi-experimental study with non-random allocation of subjects to ‘control’	Control N = 75Intervention n = 65	BDI (positive Cohen’s d indicates symptom improvement)FACES III (positive Cohen’s d: more ‘balance’ and less ‘extreme’)	Low
Rotheram-Borus et al. (2000)	Adolescent suicide attempters and their mothers in the ED	As in Rotheram-Borus et al. (1996) but collected longitudinal data	1 per family group	As in Rotheram-Borus et al. (1996) 45	As in Rotheram-Borus et al. (1996) 45	Suicide attempt rateAbove the ‘clinical range’ cut-off for BDI	Low
Wharff et al. (2012)	Adolescents with suicidal ideation or behaviour, and their families in the ED	Family-Based Crisis Intervention by trained social worker	1 per family group	Quasi-experimental design with participants included sequentially	ControlN = 150InterventionN = 67	Likelihood of admission from ED	Low
Wharff et al. (2019)	Adolescents with suicidal ideation or behaviour, and their families in the ED	As in Wharff et al. (2012)	1 per family group	Prospective randomised controlled trial	InterventionN = 68ControlN = 71	Reasons for Living Inventory for AdolescentsFamily Empowerment ScaleLikelihood of hospitalisation	Low

Studies are ordered first by similarity of intervention (‘no-suicide contract’ intervention type, and ‘specialised suicide-specific therapy’ intervention type), study type (randomised controlled trial, then quasi-experimental trial) and then by date.

ED: emergency department; NSC: no-suicide contracts; BDI: Beck’s Depression Inventory; PICO: Population, Intervention, Control, Outcomes; FACES: Family Adaptability and Cohesion Evaluation Scales.

Outcome measures

The outcome measures used to evaluate NPIs were highly variable (Tables 1 –6 have outcome measures identified for use in potential meta-analyses, and Supplementary Table 17 for all outcome measures). We identified 26 primary outcome measures from the 39 studies where a primary outcome was specified, and 71 measures in total. Outcome measures used in the four quantitative sensory modulation room (SMR) studies were consistent, using either 10 (Lloyd et al., 2014; Novak et al., 2012) or 11 (Seckman et al., 2017; West et al., 2017) point Likert scales of distress, and number/rate of seclusions (Dickson et al., 2021; Lloyd et al., 2014; Novak et al., 2012; Seckman et al., 2017; West et al., 2017). Rate of self-harm was measured across all multicomponent-individual studies (Drew, 1999; Potter et al., 2005). Both quantitative short admission studies reported on ‘days admitted to hospital’; both commented on rates of self-harm but the measures were different (Berrino et al., 2011; Westling et al., 2019). Four ‘specialised suicide-specific intervention’ studies used comparable measures of depression symptoms (Beck’s Depression Inventory [BDI] or Children's Depression Inventory [CDI]) (Ghahramanlou-Holloway et al., 2020; Rotheram-Borus et al., 1996, 2000; Wharff et al., 2012), and three measured family constructs (Family Adaption and Cohesion Scale II/III) (Rotheram-Borus et al., 1996, 2000; Wharff et al., 2012). Outcome measures were otherwise dissimilar.

Clinical, methodological and statistical results

Results and heterogeneity are discussed in each section, ordered according to category and type.

Skills-focussed intervention category

We identified one controlled and one uncontrolled quantitative trial (Table 1; Supplementary Table 7). Myhre et al. (2018) reported a statistically significant between-groups improvement on BDI-II score using Behavioural Activation. Huffman et al.’s (2014) feasibility study of ‘positive psychology’ interventions reported a high participation consent rate and statistically significant within-group improvement in ‘optimism’, but there was no control group.

Psychoeducation intervention category

We identified three quantitative and one qualitative study. Table 2 has PICO study characteristics and data, with more detail presented in Supplementary Tables 15 and 16.

Dyckman et al. (1999) found a comparative reduction in panic attack presentations following face-to-face psychoeducation with a trained specialist primarily because the presentation rate for the ‘brochure-only’ group increased by 67% compared to control (p = 0.0417).

Ward-Miller et al. (2019) found that participants’ scores on a ‘Resourcefulness Scale’ did not improve following a resourcefulness-related psychoeducation programme using puppets. Gebhardt et al. (2022) found that a single-session suicide-focused group psychoeducation programme for veterans was feasible and acceptable. Forty-nine percent reported no change in distress, but participants ‘learned useful information’. Petrakis et al. (2017) identified several themes pertaining to a psychoeducation programme regarding first episode psychosis (Supplementary Table 16), although the philosophical perspective, research methodology and reflexive influence of the authors were not clear.

Environment-focussed intervention category

Sensory modulation and SMRs intervention type

Five quantitative (Lloyd et al., 2014; Novak et al., 2012; Seckman et al., 2017; West et al., 2017) studies (all quasi-experimental) and one qualitative study (Sutton et al., 2013) were identified. All five experimental studies found showed a significant within-group reduction in ‘distress’, but between-group effect sizes were small and did not reach statistical significance (Table 3). A qualitative analysis (Sutton et al., 2013) (Supplementary Table 5) attributed this distress reduction to the soothing nature of the SMR and staff–patient interaction. Two reported reduced seclusion rates (Lloyd et al., 2014; Seckman et al., 2017), but two others found no such effect (Novak et al., 2012; West et al., 2017) (Table 3; Supplementary Table 4).

Dickson et al. (2021) reported that weighted blankets in the ED reduced anxiety, but the small sample size reduced result reliability (Supplementary Table 4).

Quiet rooms intervention type

We identified two quantitative studies (Table 3; Supplementary Table 6). Glod et al. (1994) found a statistically significant reduction in total aggression for children in modified quiet rooms (QRs). In Ray et al.’s (2017) study, it was not possible to determine whether the intervention caused the observed reduction in ‘continuous special observation’.

Special Observation intervention category

We identified only one quantitative and two qualitative studies of this clinically common bNPI (Table 4; Supplementary Tables 8 and 9). The sole quantitative study (Stewart et al., 2009) retrospectively analysed data from 16 psychiatric units over a 3-year period and found that the use of Special Observation (SO) did not influence the incidence of self-harm in the following week.

In Pitula and Cardell’s (1996) qualitative study (Table 4; Supplementary Table 9), participants reported that the presence of staff made them feel less likely to lose control over self-harm impulses. Conversly, participants also reported feeling like prisoners, and high staff turnover of those conducting SOs made them feel like a burden. A study of nurses perceptions yielded nine themes (Cleary et al., 1999) (Supplementary Table 9) which pertained to patient safety and those undertaking observations feeling unseen by doctors.

‘Multicomponent Group’ intervention category

We identified four quantitative studies of group-based multicomponent interventions (Barker and Dawson, 1998; Silverman, 2019a, 2019b; Walsh, 1993). They are presented together in Table 3 and Supplementary Table 10, but are otherwise dissimilar. No studies showed significant changes on any outcome compared to controls. There was a within-group reduction in state anxiety associated with animal-assisted therapy (Barker and Dawson, 1998) in people experiencing psychosis, and within-group reductions in state anxiety for both animal-assisted therapy and the control group in those with mood disorders. Brief music therapy yielded no significant within-group effects (Silverman, 2019a, 2019b) (Table 5).

‘Multicomponent Individual’ intervention category

No-suicide contracts or safety planning intervention type

Two quantitative studies, and one mixed methods study evaluating no-suicide contract (NSC) or safety planning (SP) were identified (Table 6; Supplementary Table 11 effect sizes).

Potter et al. (2005) found no difference in the rate of self-harm following the introduction of NSCs (where patients are asked to contract that they will seek help when distressed and agree upon useful emotion management interventions) on an inpatient ward. Drew’s (2001) retrospective file review reported significantly higher levels of self-harm in patients with an NSC to those who had no such contract which they speculated was because people who were likely to self-harm were more likely to be placed on a NSC, rather than this being an adverse effect of the contract. Bagatelas et al. (2022) evaluated the quality of SP and found that those who completed a ‘high-quality’ SP were less likely to subsequently present to ED than those with a ‘low-quality’ SP. Whether this was because of the quality of the SP itself, or whether the quality of the SP reflected levels of motivation to engage in treatment, was not explored.

Specialised suicide-specific therapies intervention type

We identified six studies. Four were quantitative studies by two different research groups evaluating two different specialised suicide-specific multicomponent interventions, and only in the ED. Both investigated child and adolescent participants and their mothers (Table 6; Supplementary Table 12 has effect sizes).

Rotheram-Borus et al. (1996, 2000) reported that both mothers and adolescents scored persistently lower than control groups on depression and suicidality measures at 3 and 18 months. Outpatient follow-up was significantly higher for adolescents in the intervention group, and for those who rated their family as higher on adaptability on the FACES III (a measure of adaptation and cohesion of the family). FACES III scores were significantly higher for the intervention group than the control group for mothers, but not for adolescents.

Wharff et al. (2012, 2019) pilot study suggested that a family-based crisis intervention reduced hospitalisation in adolescents and replicated this comparative efficacy in an RCT. The Reasons for Living Inventory and Family Empowerment Scale scores improved significantly for both intervention and control groups (p < 0.001), but with no between-group differences.

An app-based suicide-specific ED intervention investigated by Dimeff et al. (2021) was found to be feasible and acceptable. There was a statistically significant improvement between groups for the Suicide-Related Coping Scale compared to Treatment As Usual (TAU), as well as distress, agitation and coping ability, and a non-significant increase in readiness to go home, on the Safety and Imminent Distress Questionnaire (Supplementary Table 12).

Ghahramanlou-Holloway et al. (2020) conducted a feasibility study of a brief Cognitive Behavioural Therapy (CBT) intervention on an adult inpatient military ward. While 24% declined to participate, none of the participants dropped out of the study. There were no significant changes in any measures.

‘Multicomponent Individual’: short admissions

We identified two quantitative studies, five qualitative studies, and one case study (Table 7; Supplementary Tables 13 and 14). Qualitative data pertaining to patient or clinician experiential evaluation of the ‘Brief Admission’ (BA) intervention were retained. Data pertaining to intervention components were excluded.

Table 7.

PICO table of ‘Multicomponent Individual’ intervention category – ‘short admission’.

Author	Treatment population and context	Country	Intervention description from study	No. of intervention episodes	Length of intervention	Study type	No. of participants	Measure	Effect size or other reported outcomes	Risk of bias
‘Short admission’ intervention type
Helleman et al. (2016)	An individual with BPD in crisis on an Adult inpatient ward	The Netherlands	BA	95 single episodes of BA	Up to 3 days	Qualitative case study	1	Thematic	N/A	Low
Westling et al. (2019)	Individuals with BPD in crisis on an Adult inpatient ward	Sweden	BA	Negotiated separately for each individual	Up to 3 days	Randomised controlled trial	Intervention N = 62ControlN = 63	Rate of admissions to hospital over 6-month intervalRate of non-suicidal self-injury over 6-month interval	Intervention within group Cohen’s d = 1.50Control within group Cohen’s d = 1.52Within-groups Cohen’s d Control = −0.14Intervention = −0.34Between-groups Cohen’s d = 0.43(95% CI: 0.08 to 0.78)	Moderate
Berrino et al. (2011)	Individuals with BPD referred to an ED, on an adult inpatient ward	Switzerland	Admission plus individual manualised psychotherapy intervention	Not stated	Maximum 5 days	Quasi-experimental study with separate pre- and post-intervention group	N = 100 controlN = 100 intervention	Likelihood of self-harmNumber of hospitalisations	Odds ratio intervention vs control = 0.42Odds ratio intervention vs control = 0.63	Low
Nehls (1994b)	Individuals with BPD who have participated in a BHTP, on an adult inpatient ward	USA	BHTP	Not stated	48–72 hours	Qualitative	5	Thematic	N/A	Moderate
Nehls (1994a)	Inpatient and outpatient nurses caring for people who had experienced a BHTP, on an adult inpatient ward	USA	BHTP	Not stated	48–72 hours	Qualitative	13	Thematic	N/A	Moderate
Helleman et al. (2014)	Individuals with BPD who have experienced a BA intervention, on an adult inpatient ward	The Netherlands	BA	2–68	Maximum 3 days	Qualitative	17	Thematic	N/A	Low
Eckerström et al. (2019)	Nurses who work on wards that use the BA intervention, on an adult inpatient ward	Sweden	BA	N/A	Maximum 3 days	Qualitative	8	Thematic	N/A	Moderate
Eckerström et al. (2020)	Individuals with BPD who have experienced the BA intervention, on an Adult inpatient ward	Sweden	BA	1–14	Maximum 3 days	Qualitative	15	Thematic	N/A	Moderate
Lindkvist et al. (2021)	Adolescents with BPD who have experienced the BA intervention, on an inpatient ward	Sweden	BA	At least 1	3 days	Qualitative	19	Thematic	N/A	Low

Intervention type. Studies are ordered as follows: First by type of study (case report, then randomised controlled trial, then quasi-experimental); then by context (inpatient, ED); then qualitative, by year (early to late).

ED: emergency department; BA: Brief Admission; CI: confidence interval; BHTP: Brief Hospital Treatment Programmes; BPD: Borderline Personality Disorder.

The two quantitative studies were undertaken in Scandinavia. At 3-month follow-up, Berrino et al. (2011) found that a short admission with concurrent manualised psychological intervention yielded a statistically significant reduction compared to TAU, in suicide attempt repetition, mean suicide attempt day survival, hospitalisation survival and mean number of days in hospital.

Westling et al. (2019) compared a BA which is ‘an intervention allowing individuals to hospitalize themselves. Duration . . . and frequency . . . are limited’ (p. 14), and specific goals of the intervention were discussed prior to admission to TAU. Over a 6-month period, they found significant reductions in days in hospital, number of visits to ED and self-harm rates within both intervention and controls but no between-group differences, probably reflecting regression to the mean. The intervention group achieved statistically significant improvements on the World Health Organization Disability Assessment Schedule (WHODAS) II mobility domain. At 12 months, those who had a BA maintained some improvement in function, but showed no reduction in repeat hospitalisations compared to TAU.

Qualitative studies

Supplementary Table 14 shows detailed qualitative results.

Five papers included patient participants and one qualitative case study; four were conducted in Scandinavia among research teams who have common affiliations, evaluating the BA approach (Eckerström et al., 2020; Helleman et al., 2014, 2016; Lindkvist et al., 2021) (Table 3), and the other investigated Brief Hospital Treatment Programmes (BHTP) in the United States (Nehls, 1994b). Commonalities included hospital as a ‘timeout when times were tough’ (Eckerström et al., 2020), ‘respite’(Nehls, 1994b), ‘feeling safe and relieved’(Lindkvist et al., 2021), ‘time out from daily life’(Helleman et al., 2014) and ‘crisis containment’(Helleman et al., 2016).

Two papers included nurse participants. In both studies, the nurses identified ‘empowerment’ of the patient as a key outcome and emphasised that the environment needed to be ‘caring’, though the participants in Nehls’ (1994a) study identified the inherent conflict within the term ‘mandated caring’. Participants in Eckerström’s (2019) study emphasised the goal of partnership between patient and nurse.

Discussion

Principal findings

This is the first systematic review of evaluations of bNPIs for people in a crisis of mental ill-health, in EDs and inpatient units. Studies that found between-group differences on inpatient units include Behavioural Activation (improved BDI) (Myhre et al., 2018), QR (reduction in total aggression in children) (Glod et al., 1994), Safety Planning (reduction in ED readmissions as long as they were ‘high quality’) (Bagatelas et al., 2022), BAs with a manualised psychological intervention (reduced suicide attempt repetition, mean suicide attempt day survival, hospitalisation survival and mean (number of days in hospital) (Berrino et al., 2011) and BAs without a manualised psychological intervention (improvements on the WHODAS II mobility domain) (Westling et al., 2019).

In the ED setting, between-group differences were found in suicide-specific ED interventions (reduced depression and suicidality measures immediately; Dimeff et al., 2021) and at 3 and 18 months (Rotheram-Borus et al., 2000), higher outpatient follow-up rates (Rotheram-Borus et al., 2000), reduced rates of hospitalisation (Wharff et al., 2019) and face-to-face specialist psychoeducation on panic (reduced ED presentation rate) (Dyckman et al., 1999). However, common interventions such as sensory modulation, weighted blankets, psychoeducation, continuous SO, animal therapy, music therapy and ‘no suicide’ contracting failed to show between-group benefits. Very few studies were replicated, making patterns difficult to identify.

Intervention settings

A visual map of the studies (Figure 1) indicates that the bNPIs evaluated in ED and inpatient settings are different and that considerably fewer interventions have been studied in EDs. This is significant, as it highlights an untapped opportunity to initiate interventions within the ED context and seamlessly extend their treatment into inpatient settings. While it is understandable that group therapies have not been evaluated in EDs, other bNPIs such as sensory modulation and SMRs, the use of SO, and skills-focussed interventions are used in this setting have not been evaluated. Conversely, suicide-specific interventions that have some evidence of effectiveness in the ED have potential for implementation in short-stay inpatient environments. Although there may be less if any impact in reducing suicide mortality (Fox et al., 2020), there appears to be other valuable therapeutic effects. Having said that, of the four interventions tested in ED, only one was even partially performed by ED clinicians. Investigating the feasibility of ED clinicians delivering psychiatric interventions (as has been done for drug health–related intervention (Kaner et al., 2018)) is vital, given the majority of EDs do not have access to a sub-specialised psychiatric services. Finally, almost all ED interventions were studied in the United States, a possible reflection of geographically specific priorities.

Synthesis of effectiveness of brief NPIs

This review found that the evidence base for crisis-related brief NPIs is insubstantial, with such inconsistency in outcomes that it seems impossible to claim that patients are being treated in an evidence-based way. We found little evidence on which to base education, clinical practice, guidelines and policy and significant gaps for many commonly used NPIs. Table 8 summarises the discussion regarding each intervention category (Supplementary Table 18 for extended discussion).

Table 8.

Main findings and future research recommendations according to intervention category and type.

Intervention category	Intervention type	Discussion of main findings	Future research recommendations
Environment-focussed interventions	SM and SMR	SMRs alone do not reduce aggression (Novak et al., 2012; Seckman et al., 2017)SM aids relaxation, self-soothing and stress reduction (Lloyd et al., 2014; Novak et al., 2012; Seckman et al., 2017; West et al., 2017)Mechanism of action is unclear	Qualitative studies that hypothesise about expected outcomes and mechanisms of change.Larger quantitative studies to investigate utility of SMRs and QRs on inpatient wards and EDs, for whom they have efficacy and under what conditions
Environment-focussed interventions	Quiet rooms	May reduce time taken to manage agitation in children (Glod et al., 1994)Mechanism of action unclear – hypothesised to work by ‘reducing some forms of stimulation while providing alternative sensory cues’ (Glod et al., 1994: 560)
Skills-focussed interventions	Positive psychology Behavioural activation	Preliminary results suggest inpatient Skills-focussed interventions (‘positive psychology’) effectively improve a range of symptoms (Huffman et al., 2014)Theory of change may be in keeping with cognitive behavioural therapy theories (Huffman et al., 2014; Myhre et al., 2018)	Large quantitative studies that use validated outcome measures that reflect mechanism of change. Extend outcome measures to include rates of attendance at follow-up, rates of re-hospitalisation, rates of self-harm in and out of hospital. Extend interventions to other skills.
Special Observation	Special observation	Insufficient data to inform the prescription of special observation in clinical practiceSome data to potentially inform service planning (Stewart et al., 2009)Qualitative data suggest engagement with staff is helpful (Pitula and Cardell, 1996; Cleary et al., 1999)Theory of change is not elucidated in quantitative study (Stewart et al., 2009)	Qualitative studies with clinicians and service-users to establish hypotheses about expected outcomes and mechanisms of change, taking into account attachment styles, psychiatric diagnosis and symptomatologyClinically useful, robust quantitative studies
‘Multicomponent Group’ interventions	Art therapy Animal therapyMusic therapy	Mechanisms of action and expected outcomes unclear (Barker and Dawson, 1998; Silverman, 2019a, 2019b; Walsh, 1993).No statistically significant change to symptomatology	Feasibility studies on short-stay inpatient wards – intervention may not be suited to this populationQualitative studies required to identify desired and expected outcomes. Possible quantitative studies, depending on the above
‘Multicomponent Individual’ interventions: suicide-specific	No suicide contracts/safety agreements	Evidence does not suggest that inpatient no-suicide contracts/safety agreements reduce the frequency of inpatient self-harm episodes (Potter et al., 2005)High-quality inpatient safety plans may be correlated with reduced likelihood of re-presentation to ED (Gebhardt et al., 2022)	Qualitative data urgently needed to understand expected outcomes and usefulness of safety planning and no-suicide contracting on the inpatient ward, and whether/how a reduction in self-harm on inpatient wards may be achieved.More, robust, quantitative studies required. Outcome measures should include rates of self-harm, re-presentation rates.
	Specialised suicide-specific therapies	Specialised interventions in ED are feasible (Dimeff et al., 2021; Rotheram-Borus et al., 2000; Wharff et al., 2019)May reduce hospitalisation rates (Wharff et al., 2019), agitation, reported suicidality, coping, readiness to go home (Dimeff et al., 2021)Good clarity of outcomes reflects cohesion between theory of change and outcome measuresSpecialised training may be required	Quantitative studies should investigate whether these therapies may be further effective if used across the ED and inpatient settings
	Short admissions	Most frequently studied intervention, with considerable qualitative data, mostly coming out of the NetherlandsManualised inpatient interventions may reduce suicide attempt repetition, mean suicide attempt day survival and hospitalisation survival (Berrino et al., 2011)	Qualitative studies investigating aspects of a short admission that are helpful in maintaining control over self-harm impulses after discharge. Qualitative studies that consider service users, clinical staff and executive staff perspectives may inform the selection of outcome measures. Quantitative studies concentrating on effects of short admissions with and without manualised psychotherapies, on function and distress, rate of community follow-up, rate of post-discharge engagement with ‘meaningful activity’ and rate of presenting with the support of emergency services. Evidence-based long-term therapies could potentially be modified into brief, manualised, inpatient formats
Psychoeducation	Psychoeducation for various conditions	Brochure for panic in ED is ineffective in reducing ED presentations, though face-to-face contact may demonstrate efficacy in larger sample size (Dyckman et al., 1999).Group psychoeducation is feasible (Gebhardt et al., 2022).	Qualitative studies to understand service-user, family and carer psychoeducation needs. More, larger quantitative studies.

SMR: sensory modulation rooms; QR: quiet rooms; ED: emergency departments; SM: Sensory Modulation.

There are positives for bNPIs. Of the single modality intervention, there seems promising evidence for interventions that modulated the environment via sensory reduction and/or weighted blankets and positive skill approaches that need replication. More complex multicomponent suicide-specific interventions seem effective for some aspects of symptoms and coping, and had additional impact on the family and service use, and one coordinated research group has published several papers providing support for coordinated BAs.

Conversely, given the widespread use of SOs and NSCs, the lack of effectiveness demonstrated in the few evaluations is concerning, especially given existing evidence that safety planning is effective in other settings (though the intervention is yet to be subjected to meta-analysis) (Ferguson et al., 2022). Psychoeducation was thought acceptable although was ineffective, and the limited evaluations of brief group-based interventions of less ‘clinical’ interventions, e.g., using animals or music, showed no benefit. We found no studies evaluating other commonly used interventions such as the quantitative effects of SO, specific verbal de-escalation strategies or evidence-based psychotherapy modalities (such as mentalisation-based or transference-focussed psychotherapy) that could be adapted for crisis settings.

Outcome measures

Evaluations of bNPIs have used a wide range of outcome measures, with limited consistency even among studies with similar interventions, indicating an inconsistency in the alignment of bNPI objectives. This might arise from insufficient research appraisal, a fragmented research landscape or a disconnection between research and clinical aims. Outcome measures need to align with intervention’s theory of change, patients’ treatment expectations and anticipated rate of change in short-stay environments. It is imperative that researchers ask clinicians, patients, families and carers: what do we expect to change in such a short time, or be affected in the longer term from a bNPI? This might include questions about whether clinicians expect rates of self-harm and suicide to reduce, given their long-term often recurrent nature, or whether it is more appropriate to expect a shift in motivation to engage in treatment, a temporary change in mood and anxiety levels, intensity of suicidal thoughts, increased linkages in the community or a comprehensive follow-up plan.

Recognising the significance of outcome measures, the development of a shared core outcome set (Mestre Navas et al., 2022) is recommended, defining the minimum best-practice measurement requirements. In the absence of such a set, researchers can turn to resources like the COSMIN (COnsensus-based Standards for the selection of health Measurement Instruments) website or engage in collaborative design through methods like Delphi studies (as has been done in Australia (Hill et al., 2019)) or consensus statements (co-produced by stakeholders).

How do we progress from here

The identified quantitative studies predominantly had low sample sizes limiting power to show an effect, and most studies had a moderate risk of bias (high-risk studies were excluded). There are challenges in performing trials of complex interventions (Skivington et al., 2021), and in sourcing funding for interventions that are difficult to commercialise; however, the lack of organisation and consistency may also imply a de-prioritisation of EDs as treatment environments.

Across most studies, there was a lack of comprehensive discussion around theory of change. While most studies mentioned theory of change, it was not always clear whether, or how, the theory of change related to outcome measures. While this is difficult to quantify (Aggarwal et al., 2021; Mestre Navas et al., 2022), as goals and mechanisms of action become clear, outcome measures can be chosen so as to reflect change mechanisms. Ideally, valid and reliable measures should be prioritised over untested outcome measures.

As well as identifying possible theories of change, researchers and clinicians would benefit from a clearly defined set of patient groups and intervention goals. This is not straightforward. To understand which interventions work for whom, researchers may wish to group individuals into symptom-based groups such as suicidal thoughts or behaviours (O’Connor et al., 2020; Schuck et al., 2019), behavioural disturbance and psychosis. However, goals will not be clear without further qualitative exploration. Attempts have been made to define goals of treatment and research (D’Onofrio et al., 2010; Wilson et al., 2019) but this is demonstrably difficult to achieve (Wilson et al., 2019).

Instead, future studies may have to consider comparative effectiveness pragmatic trials, stepped-wedge studies, cluster randomised studies and studies where individuals are randomised to various TAU alternatives (Briffa et al., 2021). These should utilise pre-existing frameworks for developing and evaluating complex interventions (Skivington et al., 2021). Suicide-specific interventions in EDs show promise for outcomes other than suicide mortality, emphasising the value of using EDs for treatment rather than solely assessment.

Limitations of synthesis

There are several limitations to this review. For example, the PICO elements may have been too restrictive. Ideally, this review would be considered with reviews investigating interventions that continue into outpatient settings, otherwise the reader may have an incomplete understanding of interventions that may be offered in EDs and crisis settings. Useful studies may have been excluded because there is no standardised definition of ‘crisis’ in the literature, though the definition was designed in such a way as to maximise study capture. It may also have been reasonable to include papers on interventions on medical wards. In addition, the taxonomy of interventions was subjective. We did not review educational documents, cannot guarantee we appraised all the grey literature and we only evaluated English language documents. Thus, some papers were likely missed, although our approach minimised the risk of missing high-quality, publicly available articles with low bias.

Recommendations for policy and practice

The evidence base for NPIs in inpatient settings offers weak recommendations across the board, leaving clinicians with minimal evidence-based guidance upon which to base their practice or satisfy the requirements of guidelines. This may be why guideline documents generally do not stretch to recommending specific interventions (NICE, 2022; Wales, 2015). However, several interventions show promise, such as short admissions with accompanying psychotherapies, SMRs and suicide-specific interventions. Interestingly, reviews have found a similar lack of structure in the fields of managing restrictive practice (Baker et al., 2021), inpatient NPIs (Jacobsen et al., 2018) and acute severe behavioural disturbance (Weiland et al., 2017). This could be attributed to the inherent challenges of conducting research in emotionally charged and hectic environments. One significant challenge, understandably, pertains to obtaining informed consent, although relevant frameworks are available (Larsen and McMillin, 2011; Manti and Licari, 2018; Vitiello, 2008). Notably, informed consent was not thoroughly discussed in any of the papers included herein.

Primarily then, funding should be allocated to studies investigating bNPIs for those in a crisis that (1) clearly identify the need they are fulfilling; (2) are translatable into the clinical environment; (3) utilise outcome measures that are reliable, valid and appropriate to the theory of change; (4) that may replicate studies done in other treatment centres; (5) that look at the effects of starting bNPIs within the ED and continuing them to the inpatient ward, recognising the potential for the ED as a treatment setting; and (6) identify how they will manage the barriers of performing research in such a challenging clinical setting. Such funding should take into account other aspects of effective treatment for those in crisis that are not investigated here, such as structural environment and service model changes.

Conclusion

This systematic review indicates that there is great need for high-quality research into brief NPIs. The current evidence base is very limited; inconsistent and outcome measures are too varied to identify which NPIs may be best for which people. It might be argued that many NPIs such as ‘means restriction’ or ‘special observation’ are too obviously clinically required to need evidence, but all interventions can have risks and benefits that need to be weighed. To address this, researchers should define theories of change of their chosen interventions and map the outcome measures that pertain to the goals of treatment, as well as using pre-existing frameworks so that clinicians and policy makers can make better choices.

Supplemental Material

sj-docx-1-anp-10.1177_00048674231216348 – Supplemental material for The effectiveness of brief non-pharmacological interventions in emergency departments and psychiatric inpatient units for people in crisis: A systematic review and narrative synthesis

Supplemental material, sj-docx-1-anp-10.1177_00048674231216348 for The effectiveness of brief non-pharmacological interventions in emergency departments and psychiatric inpatient units for people in crisis: A systematic review and narrative synthesis by Jacqueline P Huber, Alyssa Milton, Matthew C Brewer, Louisa M Norrie, Saskia M Hartog and Nick Glozier in Australian & New Zealand Journal of Psychiatry

Footnotes

Acknowledgements

The authors would like to acknowledge Dr Zeeshan Shahnawaz for her essential role in editing this paper.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was support by the St Vincent’s Hospital AMR Clinician Buyback Grant.

Registration number

PROSPERO: CRD42021287496

ORCID iDs

Jacqueline P Huber

Nick Glozier

Supplemental material

Supplemental material for this article is available online.

References

Aggarwal

Patton

Berk

, et al. (2021) Psychosocial interventions for self-harm in low-income and middle-income countries: Systematic review and theory of change. Social Psychiatry and Psychiatric Epidemiology 56: 1729–1750.

Allen

Forster

Zealberg

, et al. (2002) Report and recommendations regarding psychiatric emergency and crisis services. A review and model program descriptions. APA Task Force on Psychiatric Emergency Services. https://www.psychiatry.org/File%20Library/Psychiatrists/Directories/Library-and-Archive/task-force-reports/tfr2002_EmergencyCrisis.pdf

Anderson

Goldsmith

Lomani

, et al. (2022) Short-stay crisis units for mental health patients on crisis care pathways: Systematic review and meta-analysis. BJPsych Open 8: e144.

Bagatelas

PHL

Hunt

Wolff

(2022) Assessing the quality of patient responses to a psychosocial intervention implemented on an adolescent psychiatric inpatient unit: Devising the safety plan quality metric. Rhode Island Medical Journal 105: 22–25.

Baker

Berzins

Canvin

, et al. (2021) Non-pharmacological interventions to reduce restrictive practices in adult mental health inpatient settings: The COMPARE systematic mapping review. Health Services and Delivery Research 9: 1–218.

Barker

Dawson

(1998) The effects of animal-assisted therapy on anxiety ratings of hospitalized psychiatric patients. Psychiatric Services 49: 797–801.

Berrino

Ohlendorf

Duriaux

, et al. (2011) Crisis intervention at the general hospital: An appropriate treatment choice for acutely suicidal borderline patients. Psychiatric Services 186: 287–292.

Briffa

Symons

Zeps

, et al. (2021) Normalising comparative effectiveness trials as clinical practice. Trials 22: 620.

Chaudhary

AMD

Memon

Dar

, et al. (2020) Suicide during transition of care: A review of targeted interventions. The Psychiatric Quarterly 91: 417–450.

10.

Cleary

Jordan

Horsfall

, et al. (1999) Suicidal patients and special observation. Journal of Psychiatric and Mental Health Nursing 6: 461–467.

11.

Coates

(2018) Service models for urgent and emergency psychiatric care an overview. Journal of Psychosocial Nursing & Mental Health Services 56: 23–30.

12.

D’Onofrio

Jauch

Jagoda

, et al. (2010) NIH roundtable on opportunities to advance research on neurologic and psychiatric emergencies. Annals of Emergency Medicine 56: 551–564.

13.

deSouza

Thode

Shrestha

, et al. (2022) Rapid tranquilization of the agitated patient in the emergency department: A systematic review and network meta-analysis. American Journal of Emergency Medicine 51: 363–373.

14.

Dickson

Gantt

Swanson

(2021) Effectiveness of the weighted blanket with psychiatric patients in the Emergency Department: A pilot study. Journal of the American Psychiatric Nurses Association 29: 307–313.

15.

Dimeff

Jobes

Koerner

, et al. (2021) Using a tablet-based app to deliver evidence-based practices for suicidal patients in the emergency department: Pilot randomized controlled trial. JMIR Mental Health 8: e23022.

16.

Drew

(1999) No-suicide contracts to prevent suicidal behavior in inpatient psychiatric settings. Journal of the American Psychiatric Nurses Association 5: 23–28.

17.

Drew

(2001) Self-harm behavior and no-suicide contracting in psychiatric inpatient settings. Archives of Psychiatric Nursing 15: 99–106.

18.

Dyckman

Rosenbaum

Hartmeyer

, et al. (1999) Effects of psychological intervention on panic attack patients in the emergency department. Psychosomatics 40: 422–427.

19.

Eckerström

Allenius

Helleman

, et al. (2019). Brief admission (BA) for patients with emotional instability and self-harm: Nurses’ perspectives – person-centred care in clinical practice. International Journal of Qualitative Studies on Health and Well-Being 14: 1667133.

20.

Eckerström

Flyckt

Carlborg

, et al. (2020) Brief admission for patients with emotional instability and self-harm: A qualitative analysis of patients’ experiences during crisis. International Journal of Mental Health Nursing 29: 962–971.

21.

Evans

Connell

Ablard

, et al. (2019) The impact of different liaison psychiatry models on the emergency department: A systematic review of the international evidence. Journal of Psychosomatic Research 119: 53–64.

22.

Ferguson

Rhodes

Loughhead

, et al. (2022) The effectiveness of the safety planning intervention for adults experiencing suicide-related distress: A systematic review. Archives of Suicide Research 26: 1022–1045.

23.

Fitz-Gerald

Takeshita

(2020) Models of Emergency Psychiatric Services That Work. Cham: Springer.

24.

Fox

Huang

Guzmán

, et al. (2020) Interventions for suicide and self-injury: A meta-analysis of randomized controlled trials across nearly 50 years of research. Psychological Bulletin 146: 1117–1145.

25.

Frank

Fawcett

Emmerson

(2005) Development of Australia’s first psychiatric emergency centre. Australasian Psychiatry 13: 266–272.

26.

Gebhardt

Ammerman

Carter

, et al. (2022). Understanding suicide: Development and pilot evaluation of a single-session inpatient psychoeducation group. Psychological Services 19: 423–430.

27.

Ghahramanlou-Holloway

LaCroix

Perera

, et al. (2020a) Inpatient psychiatric care following a suicide-related hospitalization: A pilot trial of Post-Admission Cognitive Therapy in a military medical center. General Hospital Psychiatry 63: 46–53.

28.

Glod

Teicher

Butler

, et al. (1994) Modifying quiet room design enhances calming of children and adolescents. Journal of the American Academy of Child and Adolescent Psychiatry 33: 558–566.

29.

Haack

Majid

Karpa

, et al. (2020) Adjacent psychiatric emergency services to the main emergency room. In: Fitz-Gerald

Takeshita

(eds) Models of Emergency Psychiatric Services That Work. Cham: Springer, pp. 37–49.

30.

Helleman

Goossens

Kaasenbrood

, et al. (2014) Experiences of patients with borderline personality disorder with the brief admission intervention: A phenomenological study. International Journal of Mental Health Nursing 23: 442–450.

31.

Helleman

Goossens

Kaasenbrood

, et al. (2016) Brief admissions during prolonged treatment in a case involving borderline personality disorder and posttraumatic stress disorder: Use and functions. Journal of the American Psychiatric Nurses Association 22: 215–224.

32.

Higgins

Thomas

Chandler

, et al. (2019) Cochrane Handbook for Systematic Reviews of Interventions. Hoboken, NJ: John Wiley.

33.

Hill

NTM

Shand

Torok

, et al. (2019) Development of best practice guidelines for suicide-related crisis response and aftercare in the emergency department or other acute settings: A Delphi expert consensus study. BMC Psychiatry 19: 6.

34.

Hill

Rufino

Kurian

, et al. (2021) Suicide ideation and attempts in a pediatric emergency department before and during COVID-19. Pediatrics 147: e2020029280.

35.

Holland

Jones

Vivolo -Kantor

, et al. (2021) Trends in US emergency department visits for mental health, overdose, and violence outcomes before and during the COVID-19 pandemic. JAMA Psychiatry 78: 372–379.

36.

Huffman

DuBois

Healy

, et al. (2014) Feasibility and utility of positive psychology exercises for suicidal inpatients. General Hospital Psychiatry 36: 88–94.

37.

Jacobsen

Hodkinson

Peters

, et al. (2018) A systematic scoping review of psychological therapies for psychosis within acute psychiatric in-patient settings. The British Journal of Psychiatry 213: 490–497.

38.

Johnston

Spencer

Wallis

, et al. (2019) Review article: Interventions for people presenting to emergency departments with a mental health problem: A systematic scoping review. Emergency Medicine Australasia 31: 715–729.

39.

Kaner

EFS

Beyer

Muirhead

, et al. (2018) Effectiveness of brief alcohol interventions in primary care populations. Cochrane Database of Systematic Reviews 2: CD004148.

40.

Kleiman

Nock

(2020) New directions for improving the prediction, prevention, and treatment of suicidal thoughts and behaviors among hospital patients. General Hospital Psychiatry 63: 1–4.

41.

Kleiman

Bentley

Glenn

, et al. (2022) Building on the past 50 years, not starting over: A balanced interpretation of meta-analyses, reviews, and commentaries on treatments for suicide and self-injury. General Hospital Psychiatry 74: 18–21.

42.

Larsen

McMillin

(2011) Ethical issues in the conduct of research at a multidisciplinary clinic. Seminars in Speech and Language 32: 338–346.

43.

Lenhard

(2016) Computation of effect sizes. Available at: www.psychometrica.de/effect_size.html (accessed 1 September 2023).

44.

Lindkvist

R-M

Westling

Eberhard

, et al. (2021) ‘A safe place where I am welcome to unwind when I choose to’-experiences of brief admission by self-referral for adolescents who self-harm at risk for suicide: A qualitative study. International Journal of Environmental Research and Public Health 19: 300.

45.

Lizarondo

Stern

Carrier

, et al. (2020) Mixed methods systematic reviews. In: Aromataris

Munn

(eds) JBI Manual for Evidence Synthesis. Adelaide, SA, Australia: Joanna Briggs Institute. https://synthesismanual.jbi.global

46.

Lloyd

King

Machingura

(2014) An investigation into the effectiveness of sensory modulation in reducing seclusion within an acute mental health unit. Advances in Mental Health 12: 93–100.

47.

Manti

Licari

(2018) How to obtain informed consent for research. Breathe 14: 145–152.

48.

McCabe

Garside

Backhouse

, et al. (2018) Effectiveness of brief psychological interventions for suicidal presentations: A systematic review. BMC Psychiatry 18: 1–13.

49.

Mestre Navas

Taubner

Pinheiro Mota

, et al. (2022) Theories of change and mediators of psychotherapy effectiveness in adolescents with externalising behaviours: A systematic review. Frontiers in Psychiatry 12: 730921.

50.

Mitchell

Crawford

Newham

, et al. (2020) The efficacy, safety and acceptability of emergency embedded Psychiatry Assessment and Planning Units: An evaluation of Psychiatry Assessment and Planning Units in close proximity to their associated emergency departments. The Australian and New Zealand Journal of Psychiatry 54: 609–619.

51.

MohammadiGorji

Bosch

Valipoor

, et al. (2021) Investigating the impact of healthcare environmental design on staff security: A systematic review. HERD 14: 251–272.

52.

Myhre

MØ

Strømgren

Arnesen

, et al. (2018) The feasibility of brief behavioural activation treatment for depression in a PICU: A systematic replication. Journal of Psychiatric Intensive Care 14: 15–23.

53.

National Institute for Health and Care Excellence (NICE) (2022) Self-Harm: Assessment, Management and Preventing Recurrence (NICE guideline (NG225)). London: NICE.

54.

Navas

Wells

Bartels

, et al. (2022) Patient and provider perspectives on Emergency Department care experiences among people with mental health concerns. Healthcare 10: 1297.

55.

Nehls

(1994a) Brief hospital treatment plans for persons with borderline personality disorder: Perspectives of inpatient psychiatric nurses and community mental health center clinicians. Archives of Psychiatric Nursing 8: 303–311.

56.

Nehls

(1994b) Brief hospital treatment plans: Innovations in practice and research. Issues in Mental Health Nursing 15: 1–11.

57.

Nordstrom

Berlin

Nash

, et al. (2019) Boarding of mentally ill patients in emergency departments: American Psychiatric Association resource document. Western Journal of Emergency Medicine 20: 690–695.

58.

Novak

Scanlan

McCaul

, et al. (2012) Pilot study of a sensory room in an acute inpatient psychiatric unit. Australasian Psychiatry 20: 401–406.

59.

O’Connor

McClay

Choudhry

, et al. (2020) Pilot randomized clinical trial of the Teachable Moment Brief Intervention for hospitalized suicide attempt survivors. General Hospital Psychiatry 63: 111–118.

60.

Ougrin

Tranah

Leigh

, et al. (2012) Practitioner review: Self-harm in adolescents. Journal of Child Psychology and Psychiatry, and Allied Disciplines 53: 337–350.

61.

Petrakis

Laxton

(2017). Intervening early with family members during first-episode psychosis: An evaluation of mental health nursing psychoeducation within an inpatient unit. Archives of Psychiatric Nursing 31: 48–54.

62.

Pitula

Cardell

(1996) Suicidal inpatients’ experience of constant observation. Psychiatric Services 47: 649–651.

63.

Potter

Vitale-Nolen

Dawson

(2005) Implementation of safety agreements in an acute psychiatric facility. Journal of the American Psychiatric Nurses Association 11: 144–155.

64.

Ray

Perkins

Roberts

, et al. (2017) The impact of nursing protocols on continuous special observation. Journal of the American Psychiatric Nurses Association 23: 19–27.

65.

Rosebrock

Batterham

Chen

, et al. (2022). Nonwillingness to return to the Emergency Department and nonattendance of follow-up care arrangements following an initial suicide-related presentation. Crisis 43: 442–451.

66.

Rotheram-Borus

Piacentini

Cantwell

, et al. (2000) The 18-month impact of an emergency room intervention for adolescent female suicide attempters. Journal of Consulting and Clinical Psychology 68: 1081–1093.

67.

Rotheram-Borus

Walker

Ferns

(1996) Suicidal behavior among middle-class adolescents who seek crisis services. Journal of Clinical Psychology 52: 137–143.

68.

Sampson

Wright

Dove

, et al. (2022) Psychiatric liaison service referral patterns during the UK COVID-19 pandemic: An observational study. The European Journal of Psychiatry 36: 35–42.

69.

Sara

Uesi

, et al. (2022) Growth in emergency department self-harm or suicidal ideation presentations in young people: Comparing trends before and since the COVID-19 first wave in New South Wales, Australia. Australian and New Zealand Journal of Psychiatry 57: 68–68.

70.

Schuck

Calati

Barzilay

, et al. (2019) Suicide crisis syndrome: A review of supporting evidence for a new suicide-specific diagnosis. Behavioral Sciences & the Law 37: 223–239.

71.

Seckman

Paun

Heipp

, et al. (2017) Evaluation of the use of a sensory room on an adolescent inpatient unit and its impact on restraint and seclusion prevention. Journal of Child and Adolescent Psychiatric Nursing 30: 90–97.

72.

Seymour

Chapman

Starcevic

, et al. (2020) Changing characteristics of a Psychiatric Emergency Care Centre. An eight year follow-up study. Australasian Psychiatry 28: 307–310.

73.

Silverman

(2019a) Comparing educational music therapy interventions via stages of recovery with adults in an acute care mental health setting: A cluster-randomized pilot effectiveness study. Community Mental Health Journal 55: 624–630.

74.

Silverman

(2019b) Music therapy for coping self-efficacy in an acute mental health setting: A randomized pilot study. Community Mental Health Journal 55: 615–623.

75.

Skivington

Matthews

Simpson

, et al. (2021) A new framework for developing and evaluating complex interventions: Update of Medical Research Council guidance. British Medical Journal 374: n2061.

76.

Stewart

Bowers

Warburton

(2009) Constant special observation and self-harm on acute psychiatric wards: A longitudinal analysis. General Hospital Psychiatry 31: 523–530.

77.

Sutton

Wilson

Van Kessel

, et al. (2013) Optimizing arousal to manage aggression: A pilot study of sensory modulation. International Journal of Mental Health Nursing 22: 500–511.

78.

Vitiello

(2008) Effectively obtaining informed consent for child and adolescent participation in mental health research. Ethics & Behavior 18: 182–198.

79.

Wales

(2015) Mental Health for Emergency Departments: A Reference Guide. Sydney, NSW, Australia: NSW Ministry of Health.

80.

Walsh

(1993) Future images: An art intervention with suicidal adolescents. Applied Nursing Research 6: 111–118.

81.

Ward-Miller

Marino

Casey

, et al. (2019) Use of puppets to teach resourcefulness skills to women who self-injure: An exploratory study. Archives of Psychiatric Nursing 33: 94–101.

82.

Weiland

Ivory

Hutton

(2017). Managing acute behavioural disturbances in the Emergency Department using the environment, policies and practices: A systematic review. The Western Journal of Emergency Medicine 18: 647–661.

83.

West

Melvin

McNamara

, et al. (2017) An evaluation of the use and efficacy of a sensory room within an adolescent psychiatric inpatient unit. Australian Occupational Therapy Journal 64: 253–263.

84.

Westling

Daukantaite

Liljedahl

, et al. (2019) Effect of brief admission to hospital by self-referral for individuals who self-harm and are at risk of suicide: A randomized clinical trial. JAMA Network Open 2: e195463.

85.

Wharff

Ginnis

Ross

(2012) Family-based crisis intervention with suicidal adolescents in the emergency room: A pilot study. Social Work 57: 133–143.

86.

Wharff

Ginnis

Ross

, et al. (2019) Family-based crisis intervention with suicidal adolescents: A randomized clinical trial. Pediatric Emergency Care 35: 170–175.

87.

Wilson

Moutier

Wolf

, et al. (2020) ED recommendations for suicide prevention in adults: The ICAR2E mnemonic and a systematic review of the literature. The American Journal of Emergency Medicine 38: 571–581.

88.

Wilson

Shenvi

Rives

, et al. (2019) Opportunities for research in mental health emergencies: Executive summary and methodology. The Western Journal of Emergency Medicine 20: 380–385.

89.

Zun

(2012) Pitfalls in the care of the psychiatric patient in the Emergency Department. The Journal of Emergency Medicine 43: 829–835.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.29 MB