Scolaris Content Display Scolaris Content Display

Cochrane Database of Systematic Reviews Protocol - Intervention

Organisational interventions for preventing and minimising aggression directed toward healthcare workers by patients and patient advocates

This is not the most recent version

Collapse all Expand all

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

To assess the effectiveness of organisational interventions that aim to prevent and minimise workplace aggression directed toward healthcare workers by patients and patient advocates.

Background

Description of the condition

Workplace aggression is becoming increasingly prevalent in health care, with serious consequences for both individuals and organisations (Dupré 2014; Hershcovis 2010). Because of this, in the past 15 years there has been a significant increase in the research and development of interventions to prevent and minimise workplace aggression. The International Labour Office (ILO), International Council of Nurses (ICN), World Health Organization (WHO) and Public Services International (PSI) launched a joint programme in 2000, and the NIOSH Workplace Violence Initiative: Research and Implementation on workplace violence began in 2002 (CDC 2014; Di Martino 2003). Interventions for the prevention and minimisation of workplace aggression are generally separated into three categories; education and training; organisational interventions; and workplace design.

Aggression is defined as “behaviour that results in personal injury or destruction of property and which may take the form of psychological devaluation and degradation as well as physical harm” (Bandura 1976). Workplace aggression is a serious occupational hazard that is spread across many work sectors. No workplace is immune to workplace aggression, however some are at a higher risk than others (Cooper 2002; Gillespie 2010; Hills 2014; Nowrouzi 2016). Healthcare workers function in highly stressful environments as well as caring for individuals in a physiological or psychological crisis, which results in these workers being at high risk of exposure to workplace aggression. Healthcare workers are increasingly becoming victims of aggressive behaviour from service users and other members of the public. Aggression toward healthcare workers from patients and their advocates is described as external aggression (Barling 2009; Kvas 2014; Nowrouzi 2016). Aggression in the form of bullying by co‐workers is the subject of a separate Cochrane review and therefore will not be covered in this review (Gillen 2017).

As far back as 2000, 4% of workers in the European Union reported that they had been subjected to physical aggression in the workplace from external sources (Paoli 2001). Recent studies report that up to 29% of healthcare workers have been subjected to physical aggression from patients, and 21.1% from patient advocates. Up to 95% of healthcare workers have reported being exposed to some form of aggression from patients, up to 50% from patient advocates and up to 40% from co‐workers or supervisors within a 6 to 24 month period (Alameddine 2015; Camerino 2008; Di Martino 2003; Hills 2013; Hills 2015; Merecz 2006; Nowrouzi 2016; Spector 2014).

The sources of this aggression are varied and not always clear. Common causes of aggression are distress and frustration, physiological imbalances, substance misuse and abuse, intoxication, and mental health conditions such as dementia, anxiety, schizophrenia and suicidal ideation. Communication skills, and expectations of patients and their advocates may also lead to instances of aggression (Alameddine 2015; Gillespie 2010). Aggression manifests in many different ways and can range from mild verbal expressions of frustration through to extreme acts of physical harm, or homicide (Ford 2010; Gillespie 2010).

Aggression can have profound effects on the physical and psychological health of the perpetrator and the victim as well as those who witness the behaviour (Liu 2013; NICE 2015). In addition, workplace aggression impacts on individuals, organisations and communities. At an individual level, occupational aggression can impact on physical and psychological health; from short‐term effects such as bruising or feelings of anger and fear through to long‐term effects such as depression, broken bones and permanent disability (Alameddine 2015). Beyond the health effects on individuals, occupational aggression may also affect victims' finances and careers because of the need to take time off work due to illness or injury, and reduced job satisfaction and commitment to an organisation (Dupré 2014; Hershcovis 2010). At an organisational and community level, occupational aggression: has an impact on the quality of service delivery; has economic costs in terms of the need to support and replace staff; and causes potential disruption to services to the community (Cooper 2002).

Workplace aggression is predominantly found in occupations where employees have unavoidable, direct contact with people in distress. As health care often involves the care of individuals in a physiological or psychological crisis, workplace aggression is highly prevalent (Liu 2013; Nowrouzi 2016). A distinguishing factor for healthcare environments is whether they are controlled or uncontrolled. A controlled environment refers to the predetermined work area that healthcare workers operate in, such as the emergency department or a ward in hospital. This is opposed to an uncontrolled environment which is not predetermined; rather it is an impromptu work space where the healthcare workers have little knowledge of the physical and social elements and must adapt themselves to suit the environment. Examples of healthcare workers who operate in an uncontrolled environment are paramedics, midwives and remote area nurses (Campeau 2008). Occupational aggression in health care is a highly researched phenomenon. Healthcare workers are regarded as one of the highest recipients of occupational aggression, exceeded only by those working as bartenders or in protective services (Buckley 2011; Harrell 2011; HSE 2016; Piquero 2013).

The prevalence of aggression in health care is difficult to measure accurately as the majority of workplace aggression is not reported. The low rate of reporting workplace aggression is a consequence of several factors that influence individuals in their decision. The most frequent factors associated with under‐reporting are: staff viewing workplace aggression as an unavoidable part of the job; and the belief that reporting an incident will not result in any change. In line with the normalisation process theory, workplace aggression appears to have become normal and expected in everyday practice (Alameddine 2015; Martinez 2016; May 2009; Nowrouzi 2016). A lack of understanding about which incidents merit reporting and a lack of organisational support and responsiveness might also have an impact on the decision to report or not report aggression (Barling 2009; Kvas 2014; Nowrouzi 2016).

Another complication in assessing the prevalence of workplace aggression is the lack of consistent conceptual and operational definitions. Terms such as workplace harassment, workplace violence, workplace incivility, antisocial work behaviour, bullying, harassment, psychological abuse and emotional abuse are all similar and often overlapping alternatives used to describe workplace aggression. The term 'violence' is used inconsistently. Violence is either used interchangeably with aggression or differentiated as a form of workplace aggression that involves behaviours intended to cause physical harm. In this review, workplace aggression is used to encompass all terms referring to behaviours by patients and their advocates that may result in harm to healthcare workers. Organisations and researchers need a standard format for reporting aggression to improve monitoring and prevention (Barling 2009; Cooper 2002; Dupré 2014).

The majority of the research conducted in the healthcare field relates to the nursing profession, generally using self‐reporting surveys. While this is the easiest method to identify the prevalence of workplace aggression, it has the limitations of recall bias and response bias, resulting in the conclusions being flawed. This however is still likely more reliable than organisational incident reporting systems (Hills 2013). Despite the difficulties in assessing the actual prevalence of aggression in health care, external aggression has been consistently reported as the most prevalent. Aggression directed at healthcare workers perpetrated by a patient is the most common form of occupational aggression, with 10% to 95% of studied healthcare workers reporting having experienced aggression from patients within the past 6 to 24 months. Patient advocates have been identified as the second most common source of workplace aggression, exhibiting aggression towards 20% to 50% of healthcare workers in a 6 to 12 month period (Alameddine 2015; Camerino 2008; Di Martino 2003; Hills 2013; Merecz 2006; Nowrouzi 2016; Spector 2014). Paramedics have also been studied. In the US, paramedics were found to have a homicide rate of 1.1 cases per 100,000 workers per year compared to a national average of 0.8 homicides per 100,000 workers per year (Maguire 2002). US paramedics' non‐fatal assault rate is 22 times higher than the US national average (Maguire 2004; Maguire 2005).

Description of the intervention

Interventions aimed at the prevention and minimisation of workplace aggression are generally grouped into three categories, each with a specific target: education and training; organisational interventions; and workplace design. Education and training is directed toward individual workers, teaching employees to anticipate, identify, and react to aggression in the workplace as well as to develop the knowledge and skills in organisational approaches to prevent, minimise, manage and report incidents of aggression. Organisational interventions are directed toward organisations and communities.They involve developing programs, policies, and work practices to prevent and minimise workplace aggression (Wassell 2009). These may alter either management or employee practices (Runyan 2000). Workplace design is directed toward the physical work environment and aims to identify the physical, environmental risk factors for workplace aggression, and design the workplace appropriately to reduce this risk.

This review focuses on organisational interventions and will not elaborate further on education and training or the physical environment.

How the intervention might work

The Haddon matrix is used in the public health sector as a generic tool which may be tailored to suit specific contexts for the evaluation of injuries and the identification of interventions. This matrix provides a clear method of categorising intervention types into distinct stages. These stages are; pre‐event, event, and post‐event. The other dimension of the matrix can be used to separate interventions based on who or what they are aiming to address; this may be the victim (healthcare worker), the vector (patient or their advocate) or the environment (physical and social). These three factors have been identified as the main contributors to workplace aggression directed at healthcare workers (Haddon 1980; Martinez 2016).

Pre‐event organisational interventions aim to prevent aggression from arising and de‐escalate any potentially aggressive person or situation before that aggression is directed at other people. Event stage organisational interventions aim to take control of an aggressive person or situation to prevent any further harm. Post‐event organisational interventions aim to ensure the health of those who were subject to aggression, respond appropriately to the aggressor and evaluate and respond to the event (Haddon 1980; Ramacciati 2016).

The following table is an example of the Haddon matrix applied to workplace aggression in health care. Each cell lists potential interventions for the specific stage of an incident and where the intervention will be directed. As the physical environment is the focus of a separate planned Cochrane review, we have omitted it from the matrix.

Phase

Victim/healthcare worker

Vector/patient or their advocate

Social environment

Pre‐event

Undertake steps to minimise risk of violence and aggression through improved situational awareness, sharing of information amongst healthcare workers and the early removal of potential weapons. Specific strategies may include:

  • risk assessments;

  • advance warning systems; and

  • removal of potential weapons or perpetrators.

Use signage, information pamphlets and social media to provide information to patients, friends and family to manage their expectations of the services provided and their behaviour toward healthcare workers. For instance:

  • informing patients of waiting times;

  • educating the public on the triage system; and

  • advertising visiting hours.

Create a work environment, through the provision of adequate resources and a strong regulatory framework, where healthcare workers are safe. For example:

  • mandatory police attendance in 'no‐go' zones;

  • explicit presence of security staff;

  • adequate staffing;

  • not working alone;

  • limit access to or visibility of valuables and medications; and

  • legislative protection.

Event

Take appropriate actions to deal with emergent situations that minimise violence and aggression against healthcare workers. These include:

  • fleeing the scene;

  • self defence;

  • de‐escalation; and

  • requesting help.

Have in place the policies and resources that clearly communicate that healthcare workers can take action to protect themselves. For example:

  • physically or pharmacologically restrain violent or aggressive patients; and

  • initiate prosecution against perpetrators.

Have in place the legislative framework, policies and resources that make it clear that healthcare workers are able to take a range of protective actions. Options that could be considered include:

  • security personnel attendance;

  • raising an alarm, seeking assistance, or using panic buttons;

  • police attendance.

Post‐event

Utilise organisational processes to ensure that all incidents are reported and that follow‐up services are utilised. These include:

  • report all incidents;

  • seek medical treatment, counselling and peer support as required; and

  • follow‐up of incident outcomes

Follow up perpetrators of violent or aggressive behaviour. Actions that could be considered include:

  • barring patients or their advocates who have exhibited aggressive behaviour in the past;

  • consideration of 'zero tolerance' policies for some categories of perpetrators; and

  • determining the reasons for or triggers of violent or aggressive behaviour.

Endorse a range of formal actions following appropriate investigation of violent or aggressive incidents. Due regard could be given to:

  • prosecution when appropriate;

  • balancing the value of consistent reporting and further action against the perpetrator; and

  • completion of incident investigation, recommendations for change and feedback to all relevant parties.

(Gates 2011; Gaynes 2016; Haddon 1980; Hills 2015; Ramacciati 2016; Runyan 2000; SA Health 2016; Wassell 2009).

Why it is important to do this review

This review will be one review in an intended set of three reviews, covering: education and training (Hills 2013); organisational interventions (this review); and workplace design (to be conducted). Any interventions that will be included in the education and training and workplace design reviews will be excluded from this review. Likewise, any interventions aimed at reducing aggression as a form of workplace bullying between healthcare workers will be excluded as these have already been reviewed (Gillen 2017).

Healthcare workers and managers face significant challenges in accessing appropriate and relevant research, synthesis of research results and translation of results into specific health service contexts in regard to organisational interventions for occupational violence (Gaynes 2016).

The most critical reasons why the questions being asked in this review are important, is that this preventable issue is still negatively impacting healthcare workers, their fellow workers and families, the perpetrators of aggression, the bystanders to the incident, employers, insurance companies, governments and society in general (Cooper 2002; Dupré 2014; Hershcovis 2010; Liu 2013). An ideal workplace in this context is one where incidents of aggression are incredibly rare or rarely escalate to episodes of violence. This is currently not the case and there is scope to improve outcomes through well‐designed, coordinated and targeted interventions aimed at both individual work settings and communities and society more broadly through the development of policies, legal structures, practices and activities that result in safer workplaces (Victorian Auditor‐General’s Office 2015).

Objectives

To assess the effectiveness of organisational interventions that aim to prevent and minimise workplace aggression directed toward healthcare workers by patients and patient advocates.

Methods

Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials (RCTs). In consideration of the complexity of conducting RCTs in work organisations, we will also include cluster RCTs and controlled before‐and‐after studies (CBAs). We will include studies where the control condition is no organisational intervention for the prevention and minimisation of workplace aggression, or a comparison with current practice.

Types of participants

We will include adult healthcare workers of any gender or profession who interact with patients, their advocates, or both, in any public or private healthcare facility. This will include at least: physicians and physician assistants; dentists; nurses and midwives; pharmacists; paramedics; allied health professionals (e.g. physiotherapists, occupational therapists, speech pathologists, speech and language therapists, medical imaging practitioners, oral hygienists, podiatrists, dietitians, opticians, audiologists); and healthcare support personnel (e.g. reception staff, healthcare aides or assistants, healthcare security personnel).

Types of interventions

We will include any organisational intervention to prevent and minimise verbal or physical aggression directed toward healthcare workers and their peers in their workplace from patients or their advocates.

There are many different types of interventions, aimed at different manifestations of aggression. For example interventions may be aimed at preventing aggression, e.g. informing patients of waiting times, or the explicit presence of security staff. Interventions can also be aimed at reducing the impact of aggression, for example, de‐escalation, or a protocol to call the police and leave the scene.

Below is a categorisation of possible types of interventions.

Examples of organisational policies and processes are:

  • job design: adequate staffing, reducing access to cash, avoiding lone working;

  • incident reporting: identification, immediate management, reporting (complete incident form), incident investigation, conclusion and management;

  • procedures for raising an alarm or seeking assistance, such as panic buttons, mobile phones;

  • organisational policies and procedures: legislative protection, zero tolerance policies, robust reporting systems, peer support for victims, on‐site security, mandatory police attendance in 'no‐go' zones, violence alert systems, training programs, staffing rules;

  • a policy for physical restraint use in response to behaviour that is challenging or aggressive by patients or their advocates;

  • a policy for barring patients or their advocates who have exhibited aggressive behaviour in the past; and

  • ensuring patients are kept informed, to limit frustration and misunderstandings.

We will include interventions that are mandatory or voluntary.

We will exclude organisational interventions that consist of education or training, since they are the subject of another Cochrane review (Hills 2015).

We will exclude bullying by colleagues as this is also the subject of another Cochrane review (Gillen 2017).

We will exclude interventions on physical structures and devices as these will be the focus of a separate Cochrane review.

Types of outcome measures

Primary outcomes

Number of episodes of aggression resulting in no harm or injury, psychological harm or injury, or physical harm or injury.

Secondary outcomes

  1. Personal knowledge, attitudes and skills relating to workplace aggression, including in relation to its prevention and minimisation.

  2. Adverse personal and organisational outcomes attributable to incidents of workplace aggression (e.g. leave days taken, alterations to workforce participation including changing work patterns or attrition, litigation and rehabilitation costs).

Reporting of the secondary outcomes listed here in the trial is not an inclusion criterion for the review.

Search methods for identification of studies

Electronic searches

We will conduct a systematic literature search to identify relevant published and unpublished RCTs and CBAs. We will adapt the search strategy we developed for PubMed (see Appendix 1) for use in the other electronic databases. We will not impose language restrictions. We will arrange for the translation of key sections of potentially eligible non‐English language papers or we will arrange that people who are proficient in the publications' languages fully assess them for potential inclusion in the review as necessary.

We will search the following electronic databases from inception to present:

  • Cochrane Central Register of Controlled Trials (CENTRAL) (Wiley Online Library);

  • MEDLINE (PubMed);

  • CINAHL (EBSCO);

  • Embase (embase.com);

  • PsycINFO (ProQuest);

  • NIOSHTIC (OSH‐UPDATE);

  • NIOSHTIC‐2 (OSH‐UPDATE);

  • HSELINE (OSH‐UPDATE); and

  • CISDOC (OSH‐UPDATE).

Searching other resources

We will also conduct a search of ClinicalTrials.gov (www.ClinicalTrials.gov) and the World Health Organization (WHO) trials portal (www.who.int/ictrp/en/).

We will check reference lists of all primary studies and review articles for additional references. We will contact experts in the field to identify additional unpublished materials.

Data collection and analysis

Selection of studies

Two review authors (ES, BT) will independently screen titles and abstracts of all the studies we find with our systematic search to identify studies for inclusion. The same authors will code them as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve'. We will retrieve the full‐text study reports/publications and five review authors (ES, BT, PO, BM, SB) will independently assess the full text and identify studies for inclusion. Where studies are identified as being ineligible, we will record the reason(s). We will resolve any disagreement through discussion or, if required, we will consult a sixth review author (DF). We will identify and exclude duplicates and collate multiple reports of the same study so that each study rather than each report is the unit of interest in the review. We will record the selection process in sufficient detail to complete a PRISMA study flow diagram and ’Characteristics of excluded studies’ table.

Should our systematic searches identify studies conducted by authors of this review, we will avoid conflict of interest by having all decisions concerning inclusion and exclusion made by review authors who were not involved with the study.

Data extraction and management

We will use a study‐specific data collection form for the collection of study characteristics, intervention details and outcome data (Appendix 2). All review authors will pilot this on at least one study in the review. Two review author (ES, BT) will extract the following study characteristics from included studies.

  1. Methods: study design, total duration of study, study location, study setting, withdrawals, and date of study.

  2. Participants: N, mean age or age range, sex/gender, inclusion criteria, and exclusion criteria.

  3. Interventions: description of intervention, comparison, duration, intensity, content of both intervention and control condition, and cointerventions.

  4. Outcomes: description of primary and secondary outcomes specified and collected, and at which time points reported.

  5. Notes: funding for trial, and notable conflicts of interest of trial authors.

Two review authors (ES, BT) will independently extract outcome data from included studies. We will note in the 'Characteristics of included studies' table if outcome data were not reported in a usable way. We will resolve disagreements by consensus or by involving a third and fourth review author (PO, SB). Two review authors (ES, BT) will transfer data into the Review Manager (RevMan 2014) file. We will double‐check that data are entered correctly by comparing the data presented in the systematic review with the study reports. A third review author (PO) will spot‐check study characteristics for accuracy against the trial report. Should we decide to include studies published in one or more languages in which our author team is not proficient, we will arrange for a native speaker or someone sufficiently qualified in each foreign language to fill in a data extraction form for us.

Assessment of risk of bias in included studies

Two review authors (ES, BT) will independently assess risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We will resolve any disagreements by discussion or by involving other authors (PO,DF, BM, and SB). We will assess the risk of bias according to the following domains.

  1. Random sequence generation.

  2. Allocation concealment.

  3. Blinding of participants and personnel.

  4. Blinding of outcome assessment.

  5. Incomplete outcome data.

  6. Selective outcome reporting.

  7. Other bias.

We will grade each potential source of bias as high, low or unclear and provide a quote from the study report together with a justification for our judgement in the ’Risk of bias’ table. We will summarise the 'Risk of bias' judgements across different studies for each of the domains listed. We will consider blinding separately for different key outcomes where necessary (e.g. for unblinded outcome assessment, risk of bias for all episodes of aggression may be very different than for an adverse personal outcome). Where information on risk of bias relates to unpublished data or correspondence with a trialist, we will note this in the ’Risk of bias’ table.

For CBAs, we will assess the studies against the same criteria as RCTs but report them as being at high risk of bias on both the random sequence generation and allocation sequence concealment items. We will exclude CBA studies if the groups are not reasonably comparable at baseline (Ryan 2013).

Assesment of bias in conducting the systematic review

We will conduct the review according to this published protocol and report any deviations from it in the 'Differences between protocol and review' section of the systematic review.

Measures of treatment effect

We will enter the outcome data for each study into the data tables in RevMan (RevMan 2014) to calculate the treatment effects. We will use odds ratio or risk ratio for dichotomous outcomes, and mean differences or standardised mean differences for continuous outcomes, or other type of data as reported by the authors of the studies. If only effect estimates and their 95% confidence intervals or standard errors are reported in studies, we will enter these data into RevMan using the generic inverse variance method. We will ensure that higher scores for continuous outcomes have the same meaning for the particular outcome, explain the direction to the reader and report where the directions were reversed if this was necessary. When the results cannot be entered in either way, we will describe them in the 'Characteristics of included studies' table, or enter the data into Additional tables.

Unit of analysis issues

For studies that employ a cluster‐randomised design and that report sufficient data to be included in the meta‐analysis but do not make an allowance for the design effect, we will calculate the design effect based on a fairly large assumed intracluster correlation of 0.10. We base this assumption of 0.10 being a realistic estimate by analogy on studies about implementation research (Campbell 2001). We will follow the methods stated in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) for the calculations.

Dealing with missing data

We will contact investigators or study sponsors in order to verify key study characteristics and obtain missing numerical outcome data where possible (e.g. when a study is identified as abstract only). Where this is not possible, and the missing data are thought to introduce serious bias, we will explore the impact of including such studies in the overall assessment of results by a sensitivity analysis.

If numerical outcome data are missing, such as standard deviations or correlation coefficients, and they cannot be obtained from the authors, we will calculate them from other available statistics such as P values according to the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Assessment of heterogeneity

Based on Verbeek 2012, we will assess the homogeneity of the results of all included studies based on similarity of intervention type and follow‐up.

We will differentiate between interventions as they are categorised via the Haddon matrix. We will consider occupational groups as similar. We will differentiate between physical and verbal aggression.

Where available, we will categorise follow‐up times of less than 6 months as short‐ term follow‐up, 6 months to 12 months as medium‐term follow‐up and greater than 12 months as long‐term follow‐up, and regard these as being different.

We will test for statistical heterogeneity using the I² statistic (Higgins2011), using the following as a rough guide for interpretation: 0% to 40% might not be important; 30% to 60% may represent moderate heterogeneity; 50% to 90% may represent substantial heterogeneity and 75% to 100% considerable heterogeneity. In cases of substantial heterogeneity (defined as I² ≥ 50%), we will explore the data further, including via subgroup analyses, in an attempt to explain the heterogeneity.

Assessment of reporting biases

If we are able to pool more than five trials in any single meta‐analysis, we will create and examine a funnel plot to explore possible small study biases.

Data synthesis

Using Review Manager 5 software (RevMan 2014), we will pool data and display the results in separate forest plots by primary study design (i.e. we will analyse and display RCTs and CBAs separately) and we will pool data from studies judged to be homogeneous. If more than one study provides usable data in any single comparison, we will perform a meta‐analysis. When studies are statistically heterogeneous, we will use a random‐effects model. Otherwise, we will use a fixed‐effect model. When using the random‐effects model, we will conduct a sensitivity check by using the fixed‐effect model to reveal differences in results. We will include a 95% confidence interval for all estimates.

When I² is higher than 75% we will not pool results of studies in meta‐analysis.

We will describe skewed data reported as medians and interquartile ranges narratively.

Where multiple trial arms are reported in a single trial, we will include only the relevant arms. If two comparisons are combined in the same meta‐analysis, we will halve the control group to avoid double‐counting.

Summary of findings table

We will create a 'Summary of findings' table using the following outcomes.

Number of episodes of aggression resulting in no harm or injury, psychological harm or injury, or physical harm or injury.

We will use the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of a body of evidence as it relates to the studies which contribute data to the meta‐analyses for the prespecified outcomes. We will use methods and recommendations described in Section 8.5 and Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) using GRADEpro software. We will justify all decisions to downgrade the quality of RCTs or upgrade the quality of CBAs using footnotes and we will make comments to aid readers' understanding of the review where necessary. We will also compile an additional GRADE table showing all our decisions about the quality of evidence and their justifications.

Subgroup analysis and investigation of heterogeneity

If we find a sufficient number of studies, we plan to carry out one or more of the following subgroup analyses where we identify substantial heterogeneity (I² > 50%) among included studies.

  1. Uncontrolled versus controlled work conditions.

  2. Lone working/community‐based working versus healthcare setting.

  3. Variability in aggression based on client characteristics (e.g. mental health issue, health issue, relatives or bystander issues).

Sensitivity analysis

If we find sufficient studies, we will a conduct a sensitivity analysis to test the robustness of our meta‐analysis results by omitting studies that we judge to have a high risk of bias.

When using the random‐effects model to combine data, we will conduct a sensitivity analysis by using the fixed‐effect model to reveal differences in results.

Where we cannot obtain missing numerical outcome data and we think the missing data may introduce serious bias, we will explore the impact of including such studies in the overall assessment of results by a sensitivity analysis.

Reaching conclusions

We will base our conclusions only on findings from the quantitative or narrative synthesis of included studies for this review. We will avoid making recommendations for practice based on more than just the evidence, such as values and available resources. Our implications for research will suggest priorities for future research and outline what the remaining uncertainties are in the area.