Psychosocial interventions in the emergency centre setting for preventing reinjury due to interpersonal violence

  • Protocol
  • Intervention


  • Claire van der Westhuizen,

    Corresponding author
    1. University of Cape Town, Department of Psychiatry and Mental Health, Cape Town, Western Cape, South Africa
    • Claire van der Westhuizen, Department of Psychiatry and Mental Health, University of Cape Town, J Block, Groote Schuur Hospital, Observatory, Cape Town, Western Cape, 7935, South Africa.

    Search for more papers by this author
  • Katherine Sorsdahl,

    1. University of Cape Town, Department of Psychiatry and Mental Health, Cape Town, Western Cape, South Africa
    Search for more papers by this author
  • Gail Wyatt,

    1. Semel Institute, University of California Los Angeles, Department of Psychiatry and Biobehavioral Sciences, Los Angeles, CA, USA
    Search for more papers by this author
  • John Williams,

    1. Semel Institute, University of California Los Angeles, Department of Psychiatry and Biobehavioral Sciences, Los Angeles, CA, USA
    Search for more papers by this author
  • Nandi Siegfried,

    1. University of Cape Town, Department of Psychiatry and Mental Health, Cape Town, Western Cape, South Africa
    Search for more papers by this author
  • Dan J Stein

    1. University of Cape Town, Department of Psychiatry and Mental Health, Cape Town, Western Cape, South Africa
    Search for more papers by this author


This is the protocol for a review and there is no abstract. The objectives are as follows:

To assess the effects of psychosocial interventions for preventing reinjury due to interpersonal violence in injured emergency centre patients.


Description of the condition

Injuries due to interpersonal violence make a significant contribution to morbidity and mortality worldwide. Annually, over 1.6 million people die due to violence with low- and middle-income countries (LMICs) accounting for over 90% of these mortalities (World Health Organization 2008 p. 49). The economically active portion of society is particularly vulnerable. In 2004, interpersonal violence was the sixth largest cause of death and disability globally in the age group 15 to 44 years (World Health Organization 2008 p. 58).

Mortality rates only reflect a small proportion of the problem. In many countries, non-fatal injuries place a considerable strain on the health system. Studies investigating youth violence have estimated that 20 to 40 seriously injured young people require hospital treatment for every one violence-related mortality (Krug 2002 p. 27). Furthermore, the consequences of these injuries are far-reaching and include more than just physical pain and disability.

There are several implications of violence and violence-related injuries. The mental health impact of these injuries can lead to persistent psychiatric disorders. Emotional distress and disorders such as posttraumatic stress disorder (PTSD), substance abuse and major depression have been documented after injuries due to violence, and in some cases, after witnessing violence (Resnick 1997; Golding 1999; Kilpatrick 2003; Ellsberg 2008). Further implications of violence-related injury place victims of violence at risk for poor long-term physical and mental health. For example, victims of violence are more likely to engage in behaviours such as risky sexual behavior, substance use, and unhealthy eating (Resnick 1997; McNutt 2002).

Studies from high-income countries (specifically New Zealand, the United States and Sweden) have shown that patients presenting with injuries due to violence are at increased risk of reinjury due to violence (Sims 1989; Dowd 1996; Ponzer 1996; Brooke 2006; Worrell 2006). In one US study, those with penetrating injuries were more likely to return with further penetrating injuries, than patients who first presented with a blunt injury (Brooke 2006). Data from New Zealand reveal that an assault-injured patient presenting to an emergency centre has a 39.5 times (95% confidence interval (CI): 35.8-43.5) increased risk of presenting with further injuries due to violence, when compared to the general population. Patients presenting with non-assault injuries had a 3.2 times (95% CI: 2.7-3.9) increased risk of subsequently presenting with an injury due to assault (Dowd 1996). These patients represent a high-risk group which could benefit from modification of risk factors for violence-related injury. There is a dearth of evidence from LMICs regarding violence-related injury recurrence, and interventions for victims of violence.

Risk factors such as adverse social conditions and mental health factors have been implicated in conferring vulnerability for violence-related injury (Arseneault 2000; Kessler 2001; Wood 2006; Elbogen 2009). The risk of violence and related injury in participants with a mental disorder is particularly significant when a comorbid substance use disorder is present (Arseneault 2000; Elbogen 2009). Attention to these factors would not only improve the individual's quality of life, but may be instrumental in preventing further mortality and morbidity due to violence.

A sizeable percentage of violence victims will access health care on an outpatient basis only, often being seen in emergency centres for less severe injuries and discharged directly from the emergency centre without formal hospital admission. In many LMICs, physical care (for example, suturing and analgesics) is the only care consistently provided, and systems are not in place to address the psychosocial needs of patients with violence-related injuries. In other words, interactions with the medical team tend to be limited to the primary reason for presentation (i.e. the injury). Consequently, opportunities to address individual risks and the sociocultural context for reinjury are overlooked. The question of, 'how can we prevent this from happening again?' may never be raised. The emergency interaction and setting can offer an opportunity to initiate primary prevention for re-injury.

Description of the intervention

Primary prevention of reinjury includes psychosocial interventions which are offered to patients with violence-related injuries and are performed or initiated in the emergency centre (with a contact session which is not solely for the purpose of acquiring the participants' contact information) for patients with violence-related injuries. One of the objectives of the intervention should be to decrease reinjury due to violence, including: mortality, self-reported reinjury or a repeat visit to an emergency centre for injury. The intervention should have a psychosocial component.

For the purpose of this review, psychosocial interventions in emergency centres will include those that target individual, family and relationship level risk factors for future poor health outcomes by intervening in psychological processes or social circumstances of the participant. Such interventions in the emergency centre may include education regarding risky behaviours, counselling, case management approaches or brief interventions for risky behaviours such as weapon-carrying and substance abuse. These interventions could be described as secondary prevention of violence as they are implemented for at-risk individuals in order to halt a potentially recurrent ‘condition’ or ongoing ‘disease’ process.

How the intervention might work

Patients presenting to an emergency centre with violence-related injuries are at increased risk for reinjury due to violence (Sims 1989; Dowd 1996; Ponzer 1996; Brooke 2006; Worrell 2006), as well as other negative outcomes. Psychosocial factors associated with reinjury due to violence have been identified in patients with violence-related injuries (Zatzick 2004; Cunningham 2009; O'Donnell 2009). Unfortunately, these needs are not always addressed (Zun 2003; Anixt 2012). The modification of these risk factors may be instrumental in protecting individuals from future violent episodes.

Why it is important to do this review

Violence prevention programmes are in use in a variety of settings, including community groups, schools, hospitals and criminal justice settings. Many of the hospital-based interventions target patients admitted with serious injuries due to violence (Snider 2009). The medical literature shows that these patients constitute a smaller proportion of violence victims than those seen only in an emergency centre and discharged (Wadman 2003; Alexandrescu 2009). The emergency centre visit could represent a 'teachable moment' for victims of violence who engage in risky behaviours (Johnson 2007). The visit could also be an opportunity to address other risk factors for repeat violent victimization or perpetration, such as mental health factors.

Some studies have addressed the economic benefits of violence prevention interventions in the emergency centre to reduce reinjury and the numerous consequences associated with violent injuries. Injuries due to violence are costly. In the US, the cost of violent deaths and injuries was estimated to be $37 billion in 2000, including medical care spending and lost productivity (Corso 2007). In light of the considerable cost involved, the potential cost-saving benefits cannot be ignored. Severe injuries are particularly expensive to treat, with medical costs in the US amounting to $24,353 and lost productivity costing a further $57,209 per injured person (Corso 2007). Some studies suggest that repeat victims may experience increasingly severe injuries, increasing their chances of serious injuries and death with each emergency centre visit (Sims 1989; Brooke 2006).

At the present time, it is not clear whether psychosocial interventions delivered in emergency centre settings can reduce reinjury due to violence. Furthermore, of the interventions showing promise, it would be beneficial to determine the specific components or types of interventions that are most effective, and in which group of patients these interventions are the most advantageous. Policymakers, government health departments and clinicians could apply the results of this review in implementing interventions in emergency centres which are most likely to be beneficial for individual patients, communities and national health systems.


To assess the effects of psychosocial interventions for preventing reinjury due to interpersonal violence in injured emergency centre patients.


Criteria for considering studies for this review

Types of studies

We will include all intervention studies including randomised controlled trial (RCTs), controlled clinical trials (CCTs), cluster-randomised trials, interrupted time series studies (ITS) and controlled before and after studies (CBAs).

Types of participants

We will include studies involving people presenting to an emergency centre with injuries due to interpersonal violence.

People presenting with self-inflicted injuries will not be included. If patients presenting with self-inflicted injuries are included in studies with participants injured due to interpersonal violence, we will attempt to separate the data.

Types of interventions

We will include any psychosocial intervention addressing risk factors for injury due to interpersonal violence. Psychosocial interventions will be defined as any intervention which targets psychological and/or social risk factors for interpersonal violence, rather than biological factors.

Psychological interventions may include, but are not exclusive to: counselling, cognitive behavioural therapy, motivational interviewing or psychiatry/psychology referral.

Social interventions may include: employment or educational interventions, mentoring, social support or case management.

Comparisons will be conducted between psychosocial interventions versus no intervention, and between different types of psychosocial interventions. The comparisons between intervention and no intervention may also be a comparison of effectiveness of services before and after a change in clinical practice was implemented. Thus, the effect of so-called 'upstream' policy changes may be documented.

Types of outcome measures

Primary outcomes
  • Reinjury to the study participant due to interpersonal violence, measured either by self-report or medical records

Secondary outcomes
  • Arrest/imprisonment (not necessarily conviction)

  • Self-reported verbal or physical conflict

  • Change in substance use

  • Other mental health factors, including mental disorders and psychological trauma

  • Change in employment or education enrollment status

  • Other medical or social factors

Outcomes measured at similar times during follow-up will be grouped where appropriate, for example short term (0 to 3 months), medium term (4 to 11 months), and long term (12 months and longer). Depending on the characteristics of the included studies and the duration of follow-up, outcomes measured at final follow-up across studies may be assessed.


The following definitions will be used for the purposes of this review. The World Health Organization (WHO) defines violence as “the intentional use of physical force or power, threatened or actual, against oneself, another person, or against a group or community, that either results in or has a high likelihood of resulting in injury, death, psychological harm, maldevelopment or deprivation” (WHO Global Consultation on Violence & Health 1996).

Interpersonal violence is further categorised as “violence inflicted by another individual or by a small group of individuals” and is divided into community and family violence in the WHO typology of violence (Krug 2002 p. 5-7).

Search methods for identification of studies

We will not restrict the search for trials by language or publication status.

Electronic searches

We will search the following electronic databases:

  1. Cochrane Injuries Group specialised register (latest version);

  2. Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library) (latest issue);

  3. Ovid MEDLINE(R), Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations, Ovid MEDLINE(R) Daily and Ovid OLDMEDLINE(R) (1946 to present);

  4. EMBASE Classic + EMBASE (OvidSP) (1947 to present);

  5. CINAHL Plus (EBSCOhost) (1937 to present);

  6. PsycINFO (OvidSP) (1806 to present);

  7. ISI Web of Science: Science Citation Index Expanded (SCI-EXPANDED) (1970 to present);

  8. ISI Web of Science: Conference Proceedings Citation Index- Science (CPCI-S) (1990 to present).

The search strategy for MEDLINE (OvidSP) (Appendix 1) will be adapted as necessary, for use in each of the other databases.

Searching other resources

We will search reference lists of retrieved articles for further relevant articles. Organisations and key personnel active in the fields of emergency medicine, trauma surgery, injury prevention and mental health will be contacted for relevant studies.

We will search the following conference proceedings for reports of potentially relevant studies:

  • World Health Organization Milestones Meetings: 2007, 2009 and 2011

  • World Conferences on Injury Prevention and Safety Promotion (abstracts published in the journal Injury Prevention): 2010 and 2012

  • Society for Academic Emergency Medicine annual meetings (abstracts published in the journal Academic Emergency Medicine): 2000 to 2013

Data collection and analysis

Selection of studies

Two authors (CvdW and KS) will examine abstracts independently for inclusion. If there is any disagreement, the full text of the publication will be examined. The full text articles will be reviewed independently by two authors and decisions will then be made on eligibility. Any differences will be resolved by discussion between the two authors. If no conclusion is reached, a third author (NS) will be asked to comment. Care will be taken to ensure that each included study meets the eligibility criteria and that each study is only included once. If these aspects are not clear from the publications, the study authors will be contacted for clarification.

Data extraction and management

Spreadsheet forms will be designed for the purpose of recording descriptive information, summary statistics of the outcome measures, the quality scale ratings and associated commentary. Data collection forms will be piloted by extracting data from five studies, independently by two authors. The authors will then discuss the suitability of the form and make necessary changes. Data will be entered into Review Manager (RevMan). We will contact study authors for any missing data.

Two authors (CvdW and KS) will independently collate the following information for each study:

  1. Description of the trials, including the primary researcher, the year of publication, study location and the source of funding

  2. Characteristics of the interventions, including the number of participants in the intervention and control groups, the number of total drop-outs per group as well as the elapsed time between the end of the intervention and assessment, the number (and location) of contact sessions after the initial emergency centre visit and the type of intervention eg. counselling, substance abuse assessment and referral. The duration of the entire intervention period will be recorded using four categories: initial contact session only, short term (0 to 3 months); medium term (4 to 11 months) and long term (12 months and longer)

  3. Details on whether the intervention: i) was specifically designed to be delivered or initiated in an emergency centre; ii) was adapted for delivery in an emergency centre; or iii) was designed to be delivered in any particular setting

  4. Characteristics of study methodology, including whether randomisation was employed

  5. Socio-demographics of participants

  6. Outcome measures employed (including psychometrics), and summary of continuous (means and standard deviations) and dichotomous data.

Assessment of risk of bias in included studies

We will assess the risk of bias using the 'Risk of bias' tool, described in Chapter 8 of the Cochrane Handbook (Higgins 2011). The following aspects will be judged using the tool: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting and ‘other issues’. For the individual studies, we will describe each aspect using information gleaned from publications. For non-RCTs, we will use the risk of bias tool, but allocate an unclear risk of bias to all included studies in the 'sequence generation' domain. For cluster-randomised trials, we will use the risk of bias tool, with particular attention to allocation concealment. If participants were recruited after the clusters were randomised, an unclear risk of bias will be assigned to this domain. Under the 'other issues' domain, the data analysis for the cluster-randomised trial will be considered, i.e. if the analysis was done taking the clustering into account. For CBAs and ITS, we will use the Cochrane Consumers and Communication Review Group (CCCRG) recommendations (Ryan 2013). The review author will then assign a 'low risk of bias', 'high risk of bias' or 'unclear risk of bias' to each aspect judged. Two authors (CvdW and KS) will perform the assessment and then compare results of the assessment. Disagreements will be resolved by discussion. If no conclusion is reached, a third author (NS) will be asked to comment. If the available information about the study is not sufficient to make a decision, the study authors will be contacted for further information.

As this review focuses on psychosocial interventions, it is most likely that the participant and study personnel will not be blinded to the type of intervention received and delivered. Attention will be paid to the blinding of personnel assessing participants at follow-up visits.

We will create a 'Risk of bias' summary table in RevMan to illustrate the assessment, and make a summary assessment of risk of bias across the studies.

Measures of treatment effect

Treatment effects for each study will be documented using risk ratios (RR) for dichotomous data (as odds ratios are considered to be more difficult to interpret (Deeks 2011)). Treatment effects for continuous outcomes will be reported using standardised mean differences (SMD) as we anticipate that studies will have used different measures for measuring the same outcomes. In these cases, the SMD is preferable, although this statistic assumes that the variability discovered in study results are attributable to the use of different measures and not other reasons. Therefore, attention will be paid to other sources of variability amongst participants, for example, the age of the participant (Deeks 2011). Confidence intervals of 95% will be reported for all effect estimates. For cluster-randomised trials, if the analysis accounts for clustering appropriately, we will extract a direct estimate of the effect measure. We will seek statistical advice regarding the method for this process. Methodologies which will be considered include, but are not limited to variance components analysis and a multilevel model.

Dealing with missing data

We will contact investigators for missing data, and assess the effect of the missing data on the results. The number of participants not completing the study will be recorded, along with the full number of participants recruited.

Assessment of heterogeneity

We will assess clinical heterogeneity by considering the different participant groups, measures and interventions. If the data are too diverse a narrative and/or tabular report will be considered, rather than a meta-analysis. We will seek statistical expertise. If a narrative report is presented, the data will be reported according to intervention type, with subheadings for the type of violence addressed and the age group of the participants.

We will assess statistical heterogeneity using a forest plot and visual examination of the plot. We will use the Chi2 test to establish heterogeneity. If an I2 value of above 80% is found (indicating substantial heterogeneity) (Higgins 2003) within various subgroups described below, the characteristics of the participants, interventions and outcomes, as well as methodological differences will be explored as reasons for the heterogeneity. We will decide whether performing a meta-analysis is appropriate, and will seek statistical help if we are in any doubt.

Assessment of reporting biases

If there are ten or more included studies, we will use a funnel plot to assess reporting bias, with standard error of treatment effect on the vertical axis. The funnel plot will then be assessed for symmetry. For asymmetric funnel plots, the reasons for asymmetry will be explored. Bias, heterogeneity, chance, etc will be considered as possible reasons. A symmetrical plot will be considered to imply a lower risk of bias or heterogeneity but the possibility of bias or heterogeneity will not be ruled out. If uncertainty exists, we will consult a statistician.

Data synthesis

We will analyse studies according to type of intervention and participant group. Meta-analyses will be conducted, if appropriate, as described in the Assessment of heterogeneity section. We will use a random effects meta-analysis model, as heterogeneity is expected with regard to participants, types of injuries and interventions.

Table 1 will outline the characteristics of studies which have been designed or adapted for delivery in the emergency centre setting. The differences in the characteristics of studies (e.g. cultural, language, age, gender, relationship-status and event-related components of the intervention) which have been designed or adapted for delivery in the emergency centre setting, versus those designed for delivery in any setting will be described in the Included Studies section of the full review.

Table 1. Emergency centre study design
Study IDOriginal study designAdaptation of study designReview authors' comments

Subgroup analysis and investigation of heterogeneity

We will undertake subgroup analyses if there are at least three studies in the proposed category, in order to assess the degree to which methodological differences between trials and intervention types might have systematically influenced differences observed in the primary treatment outcomes. Subgroup analyses will be performed for:

  1. whether the intervention was designed or adapted for delivery in an emergency centre versus any setting

  2. age groups (0-10 years, 11 to 18 years, 19 to 45 years and older than 45 years)

  3. types of violence

  4. types of intervention

  5. place of delivery of the intervention

  6. time elapsed between intervention and follow-up assessment

  7. length of intervention period (initial contact session only, short-, medium- or long-term, as above), and

  8. low-, middle- or high-income country, according to the World Bank classification (World Bank 2012).

Sensitivity analysis

Sensitivity analysis will be performed by allocation concealment. We will remove the high-risk studies and repeat the analysis. The results of the original analysis will then be compared with this subsequent analysis.


The UCLA/South African Trauma Training Research (Phodiso) Program provided training and salary support for Claire van der Westhuizen. We acknowledge the Cochrane Injuries Group for assistance provided.


Appendix 1. Search strategy


1. exp Psychotherapy/
2. exp Counseling/
3. (psychosocial or psycho-social).mp.
4. ((behavior* or behaviour* or family or families or cognitive or psycho*) adj3 (therapy or therapies)).ti,ab.
5. (interpersonal or psychotherap* or "problem solving" or operant* or reinforcement* or biofeedback* or "social skill" or "social skills" or "cognitive behavioral" or "discussion group" or "insight oriented" or "client centered" or counsel* or insight* or paradox* or psychoanalys* or psychodrama* or "role play" or "role playing" or transactional or befriend* or mentor* or "social support").ti,ab.
6. (psychological adj1 debrief*).ti,ab.
7. "interview*".ab,ti.
8. (discussion adj1 group*).ti,ab.
9. "motivational interview*".ab,ti.
10. (social adj1 support).ab,ti.
11. exp Interview, Psychological/
12. (Interview* adj1 (motivational or brief)).ab,ti.
13. case management.ab,ti.
14. (alcohol* adj3 intervention*).ab,ti.
15. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14
16. exp Accidents/
17. exp abdominal injuries/
18. exp Accidental falls/
19. exp Asphyxia/
20. exp Burns/
21. exp Drowning/
22. exp Esophagus/in [Injuries]
23. exp Eye Injuries/
24. exp Eye Burns/
25. exp Lacerations/
26. exp Multiple Trauma/
27. exp Near Drowning/
28. exp Rib Fractures/
29. exp Rupture/
30. exp Shock, hemorrhagic/
31. exp Shock, traumatic/
32. exp Thoracic injuries/
33. exp Trauma, nervous system/
34. exp Trauma severity indices/
35. exp Wounds, penetrating/
36. exp Wounds, nonpenetrating/
37. exp Wounds, stab/
38. exp Wounds, gunshot/
39. ((abdom* or chest or thora* or torso) adj5 (wound* or trauma* or injur* or oedema* or edema* or damag*)).ab,ti.
40. ((ingest* or drink* or eat* or ate or swallow*) adj5 (accident* or uninten* or intoxicat* or poison*)).ab,ti.
41. (accident* or injur* or trauma* or wound* or bleed* or rupture* or scald* or asphyxia* or drown* or burn* or poison* or stab* or shot* or shoot* or gun?shot or gunshot or lacerat* or shock* or electrocut*).ti.
42. ((indust* or occupation*) adj3 (accident* or injur*)).ab,ti.
43. exp Violence/pc, sn [Prevention & Control, Statistics & Numerical Data]
44. exp Domestic Violence/pc, sn [Prevention & Control, Statistics & Numerical Data]
45. alcohol drinking.ab,ti.
46. exp Alcohol Drinking/ae [Adverse Effects]
47. exp Alcohol Drinking/ep [Epidemiology]
48. exp Alcoholism/
49. exp Substance-Related Disorders/ae, ep [Adverse Effects, Epidemiology]
50. exp Aggression/de, ep [Drug Effects, Epidemiology]
51. (alcohol* adj3 (consumption or drink* or problem*)).ab,ti.
52. (alcohol* adj3 complication*).ab,ti.
53. 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 or 47 or 48 or 49 or 50 or 51 or 52
54. 15 and 53
55. exp Case Management/og [Organization & Administration]
56. exp Directive Counseling/og [Organization & Administration]
57. exp Needs Assessment/og [Organization & Administration]
58. exp "Referral and Consultation"/og, sn [Organization & Administration, Statistics & Numerical Data]
59. Emergency Service, Hospital/og, sn [Organization & Administration, Statistics & Numerical Data]
60. Trauma Centers/og, sn [Organization & Administration, Statistics & Numerical Data]
61. *Emergency Service, Hospital/
62. 55 or 56 or 57 or 58 or 59 or 60 or 61
63. 54 and 62
64. randomi?ed.ab,ti.
65. randomized controlled
66. controlled clinical
67. placebo.ab.
68. clinical trials as
69. randomly.ab.
70. trial.ti.
71. 64 or 65 or 66 or 67 or 68 or 69 or 70
72. (animals not (humans and animals)).sh.
73. 71 not 72
74. 63 and 73

Contributions of authors

All authors were involved in the development of the concept and design of this protocol. Claire van der Westhuizen and Katherine Sorsdahl developed the draft of the protocol and all the authors commented on the draft.

Declarations of interest

Dan Stein: Prof. Stein has consulted to and served on speaker's bureaus of a number of pharmaceutical companies (most recently, Biocodex, Eli-Lilly, GlaxoSmithKline, Lundbeck, Pfizer, Servier, Solvay).

All other authors: None known.