Why do acute healthcare staff behave unprofessionally towards each other and how can these behaviours be reduced? A realist review

Aim

To improve context-specific understanding of how, why and in what circumstances UBs between staff in acute healthcare settings occur and evidence of strategies implemented to mitigate, manage and prevent them.

Objectives

This review seeks to:

1. conceptualise and refine terminology: by mapping behaviours defined as unprofessional to understand differences and similarities between terms referring to UBs (e.g. incivility, bullying, microaggressions, etc.) and how these terms are used by different professional groups in acute healthcare settings;

2. develop and refine context, mechanism and outcome configurations (CMOCs): to understand the causes and contexts of UBs, the mechanisms that trigger different behaviours, and the outcomes on staff, patients and wider system of health care;

3. identify strategies designed to mitigate, manage and prevent UBs and explore how, why and in what circumstances these work and whom they benefit;

4. produce recommendations and comprehensive resources that support the tailoring, implementation, monitoring and evaluation of contextually sensitive strategies to tackle UBs and their impacts.

Search strategy

In step 2, we identified studies addressing strategies to reduce UBs among staff in acute healthcare settings by:

• Systematically searching academic databases Cumulative Index to Nursing and Allied Health Literature (CINAHL) (EBSCOhost), EMBASE Classic + EMBASE (Ovid) 1947 to 11 February 2022, and Ovid MEDLINE(R) ALL 1946 to 11 February 2022 on 14 and 15 February 2022. Search strategies comprised search terms, synonyms and index terms for: Acute care AND Healthcare staff AND UBs. Searches from existing similar reviews such as Illing et al. (2013) were consulted to aid identifying relevant search terms. 41 The National Institute for Health and Care Excellence (NICE) OECD geographic search filters were used in MEDLINE and EMBASE to exclude studies based in non-OECD countries since these were deemed not relevant to the UK NHS acute care setting and UK workplace culture. 42 The OECD filter was adapted for use in CINAHL. Limits for language and publication date were not used; however, animal, child and elder abuse studies were removed from the search.

• Conducting similar searches for trade, policy and grey literature on Google Scholar (via Harzing’s Publish or Perish software), Google, HMIC (Health Management Information Consortium) (Ovid) 1979 to November 2021, NICE Evidence Search (www.evidence.nhs.uk/), Patient Safety Network (https://psnet.ahrq.gov/) and websites, including NHS Employers and NHS Health Education England.

See Report Supplementary Material 2 for the full search strategies used across these databases and grey literature sources. The range of databases and sources was chosen to identify relevant health research and policy reports in peer-reviewed journals, trade journals and organisation reports. For example, we sourced opinion pieces likely to discuss how and why UB interventions work in trade journals (e.g. AACN Bold Voices, Nursing Times, ED Management) via CINAHL and HMIC databases. Search strategies were peer-reviewed using the Peer Review of Electronic Search Strategies method. 43 All search results were saved in EndNote X9 [Clarivate Analytics (formerly Thomson Reuters), Philadelphia, PA, USA] software and duplicates removed using University of Leeds Academic Unit of Health Economics (AUHE) guidance. 44

Inclusion and exclusion criteria

We included peer-reviewed and grey literature that helped explain how and why strategies to reduce UBs in acute care settings work and whom they benefit. The inclusion criteria shown in Table 1 were used. 1

Screening, relevancy and rigour

Screening of search results was primarily undertaken by JAA, but RA independently screened a 10% random sub-sample for quality control at title and abstract, full text and relevancy stages. An additional 50% of November 2021 Google Scholar searches were double-screened. Any disagreements were resolved by discussion between JA, RA and JM (see below). The remaining 90% of decisions at these stages were made by JAA. Title and abstract screening was performed using Rayyan.ai software (www.rayyan.ai/) and full texts were screened using Mendeley (Mendeley Ltd.). 45

Decisions regarding sources were based on the inclusion and exclusion criteria (above) and a combination of relevance (based on both the major/minor criteria below and the ability to inform programme theories, i.e. conceptual richness and depth of sources) and rigour (whether the methods used to generate the relevant data are credible and trustworthy). Assessment of rigour focused on the extent to which sources provided a detailed description of their methods and how generalisable and trustworthy their findings were based on those methods. 36

Our formal criteria for classifying the potential relevancy of sources included assessing for:

• major contribution for sources that:

  • contributed to the study aims and are conducted in an NHS context in acute care; or

  • contributed to the study aims and are conducted in an NHS context; or

  • contributed to the study aims and are conducted in contexts with similarities to the NHS (e.g. universal, publicly funded healthcare systems); and

• minor contribution for sources that:

  • were conducted in non-UK healthcare systems that are markedly different from the NHS (e.g. fee-for-service, private insurance scheme systems) but where the mechanisms causing or moderating UBs could plausibly operate in the context of those working in the NHS; or

  • contributed to the study aims and can clearly help to identify mechanisms that could plausibly operate in the context of the NHS (e.g. law, police, military).

We prioritised the sources of major relevance in relation to the above criteria. The criteria that are italicised reflect those criteria we adopted in our final review. We did not seek studies beyond health care because we reached theoretical saturation. These sources were then sorted into the above categories and assessed for their ability to inform the refinement of programme theories (theoretical relevancy and conceptual richness). Where a scarcity of ‘major contribution’ sources meant that we were unable to develop and refine aspects of the programme theory, we drew on literature from the minor relevancy criteria. In addition, we also classified documents according to their conceptual richness (thickness vs. thinness) using adapted criteria from Pearson et al. 46 These criteria reflected the usefulness of each document to the realist analysis so that we could judiciously draw on a wide range of sources in a timely manner.

The criteria for appraising conceptual richness are defined in Table 2, with examples.

Step 1 (November 2021): Identifying existing theories and scoping the literature

We included 38 documents in the initial theory-generation step, comprising 30 identified from the proposal, five from informal searches and three from the team.

Step 2 (November 2021–February 2022): Searching for evidence

The initial systematic search identified 2629 records after 99 Google Scholar search results after cross-database deduplication. Initial independent pilot screening of 54 (of 2629) records resulted in 16 papers included for full-text screening with 100% agreement. Independent screening by two reviewers (JAA and RA) of 267 (10% of 2629) records led to 72 being included for the next step. Disagreement occurred on 40 items (15%) and was resolved through discussion between JAA and RA. The remaining 2308 papers were screened by JAA against the inclusion/exclusion criteria: of these, 400 papers were selected for full-text screening.

Google and Google Scholar searches (November 2021)

Two searches were performed on Google Scholar via use of Harzing’s Publish or Perish software (each was limited to the 50 most relevant results). This was to ensure we captured relevant grey or academic literature identified by Google’s algorithms. Search strategies are presented in Report Supplementary Material 2. After duplicate removal, 99 papers remained for title and abstract screening. Papers were excluded largely because they were not related to UBs between staff (Figure 4). Sixty-three papers were selected for full-text screening and then combined with the systematic search results at the full-text screening stage. An additional 52 sources were identified through searches on Google (30) in 2021, from the project team (12) and from stakeholders (10) during 2022 (see Additional searching). In total, 603 papers were selected for full-text screening. After cross-deduplication of these various sources of literature, 537 full-text papers were eligible for screening.

FIGURE 4.

Example NVivo 12 coding structure for strategies.

Additional searching and evidence gathering (August–November 2022)

Search update (December 2022)

The above searches were repeated on 9 December 2022 (except for NICE Evidence, which was withdrawn in April 2022) to ensure our review remained up to date. The search strategies were reviewed before running the final update searches and no changes made. This identified 3890 records, which reduced to 192 records when we removed duplicates and previously identified records. After title and abstract screening, 36 papers were retained for full-text screening. Included from this update were eight papers and an additional two studies from the team (10 in total).

Total included literature (March 2023)

One hundred and forty-eight total papers were included. This included 38 papers from Step 1, 100 papers from Step 2, and 10 from our updated December 2022 search. Full search results are depicted in Figure 3.

Realist analysis

During the review, we moved iteratively between analysis of examples from the literature, refinement of programme theory and further iterative searching to test particular programme theories as required (see Figure 2). We also used the strategies listed in Table 3 to make sense of the data. 46,51 This type of analysis enabled us to understand how the most relevant and important mechanisms work in different contexts, allowing us to build more transferable CMOCs.

These strategies found in Table 3 were also drawn upon in the theory-refinement process. In this process, we compared initial theories with novel data from Step 2 literature. Novel data were compared with initial theories to see if they affirmed, refuted or altered the existing initial theory and, in these cases, the above comparison strategies were drawn upon to make a decision regarding how theories were to be refined. Where novel data leading to candidate theories were identified that did not match with any initial theory from Step 1, we formulated novel theories and tested them further using literature from within Step 2 onwards.

Descriptive analyses

To ensure that the same information is not referred to in different ways in the report, our categories of contributors and strategies – developed based on underlying common realist mechanism (i.e. how and why they work) – were also used to organise the descriptive analysis. These categories were developed in an inductive way and discussed with the team and stakeholder group to ensure rigour. Additional codes were created in NVivo 12 for data relevant to the descriptive analysis, including the experience of particular professional groups of the contributors of UB (such as nurses and doctors) and for those from minoritised communities. In Chapter 5, we report on some basic statistics, such as mean or median sample sizes for included studies. These statistics were computed in Microsoft Excel using in-built formulae.

Categorising and collapsing contributors and strategies

When developing our programme theories to underpin individual contributors and strategies, the theories became too numerous and complex; we needed to collapse these to be more manageable and reportable. We accomplished this by forming categories according to common mechanisms underpinning how contributors lead to UB, as well as how strategies may mitigate or reduce UB. These categories were initially formed by JAA before refinement through intensive discussion with RA and JM, and input from the wider team and stakeholder groups.

Identification of key dynamics

We operationalised many of the demi-regularities (or ‘semi-predictable patterns or pathways of programme functioning’ across studies) identified across studies as ‘key dynamics’36 and ‘implementation principles’ (see Chapters 7 and 8). Key dynamics were defined as contradictions within, considerations for or frequent unintended consequences of interventions/strategies. These dynamics and principles were iteratively discussed within the team before they were presented to our stakeholders to be sense-checked against their expertise. Our intention was to surface these often-implicit contradictions, which should be actively managed by those involved in implementing interventions in organisational settings.

Formatting of context, mechanism and outcome configuration in this report

Throughout this report, CMOCs are most commonly formatted as ‘if, then, because’ statements for simplicity: meaning ‘if [context], then O [outcome], because [mechanism]’. In Chapter 6, where particular strategies are discussed, we formulated the CMOCs using a method outlined by Dalkin et al. 35 This format is as follows: R (Resources introduced by the intervention) + C (Context) → M (Change in participant reasoning) = O (Outcome).

The purpose of the differing CMOC formulation in Chapter 6 is to make a greater distinction between the resource offered by the intervention (i.e. the strategy) and the context in which that strategy is delivered.

Source types

The 148 included sources encompassed 113 empirical and 45 non-empirical sources (Table 4). The largest source type was acute care intervention papers (n = 42) (Figure 5), all identified in Step 2 onwards. These interventions, the characteristics of their evaluations and their components are discussed in Chapter 5. Full details on all 148 included sources – including country, healthcare setting, samples, etc. – can be found in Appendix 3, Table 25.

FIGURE 5.

Pie chart depicting distribution of included literature and to which step of the project it corresponds.

Healthcare settings

Included sources focus predominantly on acute healthcare settings, as per our protocol, comprising 37% of included sources (Table 5). We also included ambiguous healthcare settings (e.g. sources that referred to simply ‘bullying in health care’) or sources that encompassed multiple healthcare settings (general healthcare settings, 38.5%). We also included several sources focusing on a setting with medical professionals who were still in education or training. These comprised 5.4% of sources.

Countries of focus

We collated information regarding the country on which the source content focused, not the country in which the source was published. Our included sources were predominantly focused on the USA or UK, together comprising over 52% of sources. A further 24.3% had a focus in no specific geographical region (Table 6).

Dimensions of terminology

Some terms – such as ‘unprofessional behaviour’ – are used in more ambiguous ways and can include ‘poor or disrespectful communication, irresponsible behavior, inappropriate care, and lack of professional integrity’. 53 As such, this term encompasses all behaviours that compromise a professional environment, from more ‘active’ behaviours (e.g. being rude to a co-worker) to those that can be considered more passive (e.g. being late to work). Disruptive behaviour is another similar term that was used by included sources but typically focuses more on those behaviours that compromise patient safety.

As well as terms that encompass more passive types of behaviour, there are also those that are more ‘active’, targeted and intended to cause distress or harm. These include bullying, harassment, disrespect, rudeness, conflict, and those that are actively harmful. Lateral violence is one example; it has been defined as ‘any repetitive behaviour among peers that is considered offensive, abusive, or intimidating by the target’. 54 This suggests that lateral violence typically occurs between individuals in the same group (e.g. frontline nurses) and highlights another dimension of included UB terms: whether or not they are targeted.

While lateral violence involves staff on the same hierarchical level, other forms of UB may require a vertical hierarchy. This is typical for bullying, for example. Included articles suggested that bullying was an interpersonal form of UB that was repeated and often from a person higher in an organisational hierarchy towards a person lower in the hierarchy: ‘bullying encompassed a range of disruptive, repetitive, and ineffective behaviors, such as criticism and humiliation, negative acts perpetrated by an individual in a position of power intended to cause fear in a targeted individual’. 55 This adds two further dimensions to UBs: frequency and requirement of a hierarchy.

Unprofessional behaviours can also be more insidious and invisible. While such behaviours are difficult to target or measure, they can also be highly targeted in nature. Such behaviours can include undermining, which is defined as ‘conduct that subverts, weakens or wears away a person’s confidence, and may occur when one practitioner intentionally or unintentionally erodes another practitioner’s reputation or intentionally seeks to turn others against them’. 56 As such, the behaviour may be invisible from the target’s perspective and can be difficult to identify from an organisation’s perspective. This adds the dimension of visibility to UBs.

Mapping dimensions of unprofessional behaviours

The above results identified that UBs can: be general or more specific in nature, be more or less targeted, require an organisational hierarchy or not, inherently be frequent (i.e. must occur more than once) as per its definition, and be more or less visible to both the organisation and the target. Figure 6 outlines our typology of UB terms and their location within the various dimensions. Terms such as ‘deviant’, ‘disruptive’ and ‘unacceptable behaviour’, which include non-targeted passive behaviours, sit at the top-left of Figure 6 and none are inherently frequent by definition (i.e. they can all be one-off events). Inherently frequent behaviours include bullying/mobbing, terrorisation, victimisation and injustice. Some of these behaviours – such as bullying – also require an organisational power hierarchy (according to most definitions), whereas micromanagement also requires a hierarchy but does not need to be frequent.

Most behaviours on which we focus in this review are targeted interpersonal behaviours, such as incivility, conflict, harassment, aggression, rudeness, microaggressions and disrespect. Some are also specifically targeted towards a person or entire group but are less visible. Such behaviours include discrimination, scapegoating and ostracism (see Figure 6). We considered a dimension of severity but these were fraught with complexity and are likely to be subjective, related to recipients’ perceptions. However, we acknowledge that some types of UB are broadly considered to be more ‘severe’ than others, that is, physical assault is worse than rudeness.

Conflicting definitions

Included sources often used terms in conflicting ways. Bullying, it seems, is often used as a catch-all term for UBs of all kinds – perhaps because it has been in use in the literature for a longer period. However, ‘there is no single, universal definition of workplace bullying either nationally or internationally’. 57 While many definitions suggested a hierarchy or power imbalance was essential for bullying, others did not. One definition stated that ‘[bullying has] been used to explain aggression between colleagues who are on the same level within the organizational hierarchy and who, because of their (supposed) low personal self-esteem and poor group identity, direct abusive behaviour toward each other’. 58 Evidently, this definition does not suggest that a hierarchy must be present. For the purposes of our review – and since most definitions state that bullying occurs within a hierarchy – we have included it in the part of our typology that requires hierarchy. Without the hierarchy component, it is not clear how the definitions of bullying and harassment differ as both are frequent forms of persistent UB.

Similarly, while it might be expected that ‘lateral violence’ occurs at the same level, there was disagreement on this in the literature. One included source suggested that: ‘Terms such as horizontal violence and lateral violence suggest the perpetrator is a nurse colleague of equal status, but this is not always the case. It might be a person in a higher position – or it might not even be a nurse’. 59 Table 8 depicts examples of selected terms where conflicting definitions were identified.

How conflicting terminology impacts staff in practice

Included sources emphasised the importance of terminology and that staff understanding forms of UB such as bullying, microaggressions etc. – including their definitions and appearance – is essential to staff understanding their experience of UB in the workplace and for being able to interpret when it is appropriate to speak up. 65–67 One source stated ‘the absence of a comprehensive descriptive framework capturing and cataloguing those behaviours make identification, seeking assistance and intervention difficult’. 68 Included literature highlighted that the lack of uniform definitions can result in different interpretations of UB by staff members and from different perspectives (e.g. individual vs. organisational), which creates an atmosphere of confusion and inhibits speaking up.

Lack of uniform definitions also impacts the ability to assess prevalence of UB or identify where and how it is occurring, which can make implementing strategies more difficult. For example, differing definitions of what constitutes UB could impact ability to understand its prevalence, reduce the likelihood of speaking up and impact ability to intervene effectively. One study showed a lack of a uniform definition and clarity regarding bullying in the nursing profession meant that: ‘those who are exposed to such behaviour, including new graduate nurses, senior nurses, and nurse unit managers would have reported based upon their own understanding of these behaviours’. 58

Evolution of literature focus

One general trend noted in the literature over time was the articulation of more specific forms of UB, indicating that greater attention was paid to the experiences of minoritised communities – through the study of microaggressions or racism, for instance. All sources (n = 7) discussing microaggressions and racism more broadly were published from 2019 onwards, which is reflective of the more recent societal focus on racism and discrimination. 69–75 The same can also be said for literature relating to sexism (n = 9), which begins in 201362 and is more prevalent from 2016 onwards. 11,32,62,76–80

No such trends were observed for more widespread and generally used terms such as ‘bullying’ or ‘harassment’, as these have always been used frequently. While the greater focus on issues affecting groups with protected characteristics is encouraging, much more needs to be done to help address the greater burden these groups face from the impact of UB in the workplace.

Terms not included: ‘professionalism’, ‘civility’ and ‘other behaviours’

We also wanted to highlight the many adjacent behaviours that are not included in our review. These include those behaviours that may be the ‘inverse’ of UB, including behaviours and terms such as ‘civility’ and ‘professionalism’. It is important to highlight that professionalism may be more than the simple absence of UB, such as maintaining appropriate appearance, upholding patient confidentiality, applying clinical skills according to standards, and interacting honestly with patients. 81 On the other hand, some behaviours may be unprofessional from the perspective of an organisation – but not from that of colleagues. One example is the use of workarounds employed by healthcare staff that may deviate from organisational procedures perceived as barriers (and hence may be unprofessional from the perspective of the organisation) but that can improve patient safety. 82 Such nuanced cases are beyond the remit of this review; only interpersonal forms of UB are included.

The ‘bad apple’ approach

We noted a focus in the literature on individual and personal characteristics as contributors to UBs. One source, for example, discussed ‘maladaptive personality traits’ – such as being ‘paranoid, narcissistic, passive-aggressive and borderline types’ – as well as describing instigators of UB as having ‘poorly controlled anger’ and experiencing a ‘spillover of home problems’. 91 These individuals are determined by the literature as having a greater proclivity to engage in UB than others and are often referred to as ‘bad apples’. 92 The organisational response to such individuals is often to weed them out and try to discipline them. Over time, the hope is that this may lead to culture change, as a critical mass of instigators have their behaviour addressed.

However, other literature in our review argued that a focus on individual characteristics is often used by organisations as a ‘get out of jail free’ card that enables abrogation of responsibility and accountability to implement wider cultural change or strategies targeting UB. 61,93,94 Moreover, evidence regarding individual-level contributors – such as personality types, gender or professional group – is often very mixed. For example, in terms of professional group, some sources find surgeons to be more frequent instigators of UB towards nurses because of inherent power dynamics, whereas other sources find nurses more frequently uncivil towards one another as a result of negative self-esteem, competition and fear. 32 Additionally, possessing certain personality traits does not guarantee that someone will behave poorly, which makes the effectiveness of understanding such contributors questionable. 95

Therefore, it may be more productive for organisations to focus on modifiable factors that are targetable by interventions in order to reduce UB. This can include improving working conditions, improving climate and culture, fostering a psychologically safe culture and eliminating barriers to providing high-quality care. 96 In this manner, an organisation can create an environment that is least enabling of UB, regardless of the individual staff working at their organisation. Mannion et al. referred to this as addressing problems at the level of bad cellars (organisations), bad barrels (health systems) and bad orchards (professions), rather than bad apples. 92 Professional accountability programmes such as Ethos in Australia and Vanderbilt in the USA show promise in fostering this kind of culture change. 97,98 As such, this report and chapter will focus on factors that can be modified by interventions and not focus on what makes someone a ‘bad apple’ or a less resilient target.

Category 1: Workplace disempowerment

Workplace disempowerment can be caused by (1) organisational hierarchies, (2) physical environment and (3) unfairness. 70,99–101 How these elements contribute to UB will be explained further, aided by presentation of CMOCs throughout this section.

Category 2: Organisational confusion, uncertainty and stress

When staff experience organisational confusion, uncertainty and stress, these factors contribute to increased instances and experiences of UB. The following four headings demonstrate how this happens, showing that when staff experience a lack of control in their day-to-day work, they encounter challenges in building relationships that, in turn, increase conflict: (1) organisational change, (2) a lack of resources and high job demands, (3) a culture of pressure and (4) a lack of role clarity.

Category 3: Job and organisation designs that inhibit social connection

When job or organisational designs inhibit the ability of staff to connect with one another, social isolation and siloed working occur. This can increase the likelihood of UB occurring as well as reduce one’s ability to cope with UB, if present, through: (1) a lack of social support, (2) shift or agency working and (3) reduced ability to communicate effectively.

Reduced ability to communicate effectively

Connected to social support and many other contributors of UB, the ability to communicate effectively is also crucial. As highlighted earlier, several factors can reduce the ability to communicate effectively, including a culture of pressure or intimidation and high job demands. As such, an inability to communicate effectively can form vicious circles that increase the impact of UB on staff and directly exacerbate conflicts. This is highlighted in the following quote, in which a cycle between communication and frustration is outlined:

a lack of communication between the physician and the nurse can result in stress for the nurse […], and conflict between nurses and physicians and between nurses results in feelings of anger and frustration. 110

This causal chain is depicted in Figure 7, whereby working in a high-pressure environment, for example (C), can lead to a reduced ability to communicate effectively (M4) and, therefore, to greater frustration (O1/M5) – which can directly lead to conflict and escalating UB (O2). One study interviewing hospital staff reported that a high-pressured clinical environment could also become associated with feeling psychologically unsafe:

…I felt unsafe in that environment… In the theatre environment at this hospital, I have been sexually harassed verbally and bullied by senior staff …. 77

Additionally, an inability to communicate begets further UB as it undermines a sense of social support. 109 Sources highlighted that ‘simple gestures such as open, honest, transparent communication go a long way to build rapport with workers’. 114 This is reflected in CMOC 10. Issues with communication quality can also cause clinical information to go uncommunicated and can allow medical mistakes to go unchallenged, affecting patient safety (see Figure 12). 14

FIGURE 12.

Diagram depicting programme theory of the causal chain from UB to reduced patient care quality and safety. ‘M/O’ = mechanism/outcome (depending on component of causal chain which is in focus).

CMOC 10. If staff work in a high-pressure environment or in a culture of intimidation (C), then this can lead to reduced ability to build social connections (M16/O1), a reduced sense of social norms (M15/O2) and a reduced sense of social support (M17/O3) because there is a reduced sense of psychological safety and ability to communicate effectively (M14).

Category 4: Harmful work cultures

Leadership is a key contributor to how harmful work cultures come into being and persist. Behaviours exhibited by those in senior positions enable, model or tacitly permit UB to continue in their organisations. We articulate this in the following ways: (1) complicit and permissive leadership, (2) negative role-modelling and prior learned behaviour, (3) an authoritarian culture, (4) a lack of organisational accountability, (5) organisational deafness and (6) cliques.

Use of strategies across intervention designs

There were differences between intervention types and the strategies used, outlined as follows:

• Single-session interventions drew predominantly on awareness-raising strategies such as education about UBs (n = 2). In 11 cases, this was also combined with role-playing and other activities intended to enhance and practise the ability to speak up and challenge UB in the moment.

• Multisession interventions typically drew on similar strategies but often diversified the approach to include strategies aimed at increasing the ability to speak up or improving the ability to work together, such as assertiveness training (n = 1), problem-based learning (n = 1), conflict-management training (=1) group writing or journalling (n=2), and communication training (n = 1). However, most still drew on a mix of cognitive rehearsal/role-playing and education (n = 4).

• Combined sessions with other activity interventions often sought to (1) have a training or education session to, for example, improve knowledge and awareness of UB and ability to speak up, alongside (2) non-session activities, such as the addition of a code of conduct (n = 6).

• Professional accountability interventions often including a reporting and escalation system. These interventions typically combined this with training to enhance speaking up and role-modelling by leadership (in the case of Ethos97,127) or incorporated championing – that is, encouraging individuals to role-model and espouse the benefits of the intervention (in the case of Vanderbilt interventions53,93,98,143,147,151).

• Structured culture-change interventions tended to utilise a case-by-case or bespoke response to organisational needs, including (1) action-planning to assess which strategies to implement and (2) surveys to understand the initial landscape of UB. Strategies used within these interventions included training on assertiveness, communication and conflict resolution, as well as for leaders and other social support-building strategies.

Use of theoretical frameworks and logic models

We extracted information regarding use of theoretical frameworks and logic models where it was explicitly mentioned. Descriptions of each theory are summarised in Table 11.

Five interventions used organisational theories, including Model of Improvement by Langley,141 Kanter’s theory of structural power in organisations,155 a National Center for Organisation Development practice model,146 a Three-Level Intervention Model by Longo138 and Kilmann’s Model for transforming organisations. 150

Thirteen interventions drew on psychological theories, including Social Learning Theory by Bandura,137,145,158 Oppressed Group Behaviour,104 ‘Cognitive Learning Theories’,67 a ‘Six-Step approach to curricular development’,72 a theory of Non-Violent Communication,61,156 Einarsen’s predisposing factors for workplace bullying,134 Theory of Social Interactions at Work,154 Structurational Divergence theory107 and Transformative Learning Theory,152 and a final paper incorporated Harmonising Nursing Communication Theory. 64

In terms of intervention types:

• Three (out of 13) single-session interventions drew on a psychological theory.

• Seven (out of 11) of the multisession interventions drew on a psychological theory.

• Four (out of six) of the combined sessions drew on psychological theories (n = 2) and organisational theories (n = 2) (Table 11).

• All of the professional accountability interventions had entirely unknown theoretical frameworks (n = 8).

• Four structured culture-change interventions drew on organisational development (OD) and organisational theories, with one drawing on a psychological theory; thus, all of this intervention type drew on some kind of theoretical framework.

As such, these results indicate that use of theoretical frameworks is most common in structured culture-change interventions and multisession interventions.

Use of theory in included studies was either weakly integrated or not adequately reported (Figure 16). In those studies that did draw on a theoretical framework, intervention developers generally used theory to understand only one aspect of the causal chain linking intervention components to a reduction in UB. For example, interventions either had a focus on ‘how people learn’ (e.g. Social Learning Theory) or ‘how conflict arises’ (e.g. Structurational Divergence Theory). No included intervention depicted a comprehensive logic model explaining the entire causal chain from intervention components to a decrease or reduction in impact of UB.

FIGURE 16.

Pie chart depicting use of theory in included interventions.

Intervention participant backgrounds

In terms of the professional background of participants, in 12 studies the samples were nurses,61,62,137,141,142,145,148,149,153,156–158 nursing students were sampled in three studies104,139,144 and ‘newly enrolled nurses’ in one further study. 67 One study included only doctors. 71 Another study drew on ‘nurses, students and faculty’. 126 A further 11 studies drew on a range of staff members. 54,64–66,72,93,107,134,146,150,152 Three studies drew on organisational ‘units’,63,154,155 seven on entire hospitals, multiple hospitals or healthcare organisations53,97,127,138,143,147,151 and three were unclear. 98,103,125 Unfortunately, it was not possible to combine and include professional groups for each intervention type in Table 12 due to heterogeneity and some studies including multiple professional groups; however, these data are tabulated for each study in Appendix 6, Table 28.

Table 12 depicts the average (mean and median) number of participants according to intervention type. Although numbers are limited, it shows that numbers for training interventions were similar, while efforts in structured culture-change interventions were typically broader in scope, reaching much greater numbers of participants. Some studies used an approach of intervening in entire units or sections of hospitals, with limited information provided with respect to the number of employees who received the strategies (n = 9) (Table 12); this may be due to the nature of the intervention (e.g. implementing a reporting system). Other interventions (n = 28) were delivered to a specific number of people, ranging from five participants to 5192. A small number of interventions had no defined number of organisational units or participants (n = 5). Interventions delivered to entire units or organisations may theoretically be more effective due to their breadth of coverage.

Duration of interventions

Many studies reporting durations of sessions gave a range of, for example, 60–90 minutes; this made it difficult to calculate averages. For example, Armstrong141 used a CREW intervention lasting 4 weeks with one 20- to 30-minute meeting per week, and Jenkins et al. 144 drew on a journal club intervention that had monthly 6-hour sessions for 6 months. Nonetheless, we have synthesised intervention durations where possible according to our identified intervention types.

In terms of intervention types, for single and multisession interventions we calculated total ‘contact time’ in minutes, rather than study duration, according to when outcome measures were delivered. This is because study duration according to evaluation timing was too heterogeneous and often not reported. Where possible, we calculated study contact time by multiplying session length by number of sessions (Table 13).

• Thirteen of the included studies relied on single ‘sessions’ of 2.5-hour duration or less (see Appendix 6, Table 28). Eleven out of 13 single-session interventions reported session duration. The mean contact time of single-session interventions was 93.14 (SD 27.9) minutes with a median of 120.

• Ten out of 11 multisession interventions reported session duration. For multisession interventions the mean time was 747.91 (SD 520.0) minutes, with a median of 750 minutes. Ten studies relied on use of multiple sessions or sessions lasting longer than 2.5 hours. This shows the substantially greater contact time in multisession interventions, which could have given them greater opportunity to change behaviour. We can speculate that this may be one reason why such interventions report marginally greater effectiveness. The minimum contact time for multisession interventions was 240 minutes; this was greater than the maximum time in single-session interventions, which stood at 120 minutes.

• The combined-session interventions were too heterogeneous to quantify and four out of six of these did not provide information on duration, while one comprised 8 weeks61 and another 3 weeks. 134

• For professional accountability interventions and structured culture-change interventions, we were unable to calculate overall duration. Since these do not use ‘sessions’, their components may be engaged in by participants for multiple months or years. For professional accountability interventions, seven out of eight reported length in years. Of these, the mean was 2.56 (SD 1.20) years, with a median of 2.

• For the structured culture-change interventions, these involved the CREW interventions; two of these were 6 months in duration,154,155 with one being only 1 month (see Appendix 6, Table 28). A final CREW intervention was unclear. 146

Table 13 depicts intervention durations or contact times where we were able to homogenise these.

Effectiveness

Most studies evaluated an intervention and assessed some measures of effectiveness (Table 14). Thirteen of the 42 studies were descriptive or examined only implementation or feasibility issues (see Table 10). Of the 29 studies that assessed intervention effectiveness, the majority (n = 23) reported positive results, while six studies did not report statistically significant results – and three even reported a negative result. 54,62,66,134,153,158 The most common reason for these ‘negative’ results was the use of education strategies, which led people to become more active in reporting UB; this, in turn, led to an increase in reports after the study when compared to baseline. 66,134,158

In terms of intervention types reporting themselves as effective (where measured):

• nine out of 13 single-session interventions

• seven out of eight multisession interventions

• two out of three combined-session interventions (although one did not report statistical significance150)

• all (one out of one) professional accountability interventions

• all (n = 4) structured culture-change interventions.

Figure 17 depicts the number of studies that self-reported effectiveness. However, some, for example, simply did assertiveness training, measured an increase in assertiveness and considered that a success. 157

FIGURE 17.

Pie chart depicting number of studies reporting ‘positive’ vs. ‘negative’ results, as well as those that did not measure effectiveness.

Outcome measures

Studies used a wide range of outcome measures; the most common was change in the prevalence of UB following implementation of the intervention. Overall, 23 studies measured prevalence. Common tools used to measure experience of UB in studies included the Negative Acts Questionnaire – Revised (NAQ-R) (n = 5), the Nursing Incivility Scale (n = 3)159 and the CREW Civility Scale146 (n = 2).

One study evaluating the CREW intervention used a range of measurement tools, including the CREW Civility Scale, Effort-Reward Imbalance Questionnaire, trust in management using an Interpersonal Trust at Work Scale, Turnover Intentions and Professional Effectiveness, among others. 154 Other studies that relied upon, for instance, assertiveness training, measured assertiveness with the Rathus Assertiveness Inventory as their primary outcome. 157 Several studies relied on qualitative measures primarily, such as interviews. 93,104,143 Other studies – primarily ones evaluating the Vanderbilt or Ethos interventions – drew on statistics from their reporting systems to determine its effects. 97,151 Appendix 5, Table 28, outlines the full range of outcome measures in included studies. No studies assessed improvements to patient safety – perhaps due to their limited intervention durations or difficulty in attributing changes in such outcomes to their intervention.

Intervention design responsiveness to context

Interventions may involve single or multiple strategies. For example, most of the classroom/single-session interventions draw upon single strategies. Interventions with multiple sessions may be bundle interventions where all strategies are implemented in every context. Indeed, most interventions drew on pre-planned study designs that were fixed (i.e. their implementation could not adapt or change during the intervention period). While this provided an easier way to evaluate the intervention, the trade-off was adaptability to unforeseen events that may occur during implementation. Those using a more responsive design included structured culture-change interventions and professional accountability interventions. These approaches tended to take an organisation-wide approach, relying on a repertoire of strategies that could be implemented on-demand depending on the type of UB encountered or setting. As such, intervention content could change and flex depending on the situation, which acknowledges that organisations are changing and dynamic systems. 146 Structured culture-change interventions such as CREW,141,146,154,155 for example, scheduled regular meetings to survey the organisational context and deliver different strategies depending on what factors were deemed to be contributing to unprofessionalism at any given time or in certain units. Osatuke et al. 146 set this out in their paper outlining the CREW intervention, stating that ‘for the intervention period, the VHA organizations commit to giving time, attention, and support to having regular (weekly) workgroup-level conversations about civility’. This builds momentum and a scaffolding around which other strategies can be implemented. The intervention can vary according to its context and the authors highlight that ‘across the sites, the specifics of in-group process and follow-ups are allowed to vary greatly, and as a result, the interventions become driven by responsiveness to local needs and local culture-based civility definitions’. 146 Taking this responsive approach – particularly with structured culture-change interventions – is purported to be integral to the success of these interventions.

Professional accountability interventions, such as Ethos and Vanderbilt, rely on a reporting and escalation system at their core. However, they also implement other strategies, including capability training and championing. 97 While professional accountability interventions such as Ethos are not as flexible as CREW, they do enable a variety of different escalating organisational responses depending on specifics of individual cases of UB. 127

Interventions outside the scope of this report

Our review and search strategy focused on identifying interventions specific to UB. We did not intentionally search for those seeking to improve civility and professionalism alone. Additionally, there may be many interventions that occur in practice but are not published in the grey or academic literature. Lastly, we identified some simulation studies investigating cognitive bias modification (CBM), a technique that may limit the impact of UB on ability to report medical errors. These were not included, as simulation studies were not interventions taking place in actual healthcare practice. Full discussion of these three categories of interventions outside the scope of this report is presented in Appendix 7.

Summarising intervention findings with context, mechanism and outcome configuration

In keeping with realist methodology, we present our overall findings with respect to different intervention types with CMOCs in Table 15.

Evaluated versus unevaluated strategies

Throughout this chapter, we have made distinct whether strategies are evaluated or unevaluated, where unevaluated means not yet assessed for feasibility or effectiveness. We have separated discussion of these throughout this chapter to increase our ability to discuss whether evaluated strategies were found to be effective. We have explicitly noted all strategies we identified, as well as those that have been evaluated in some form as part of an intervention in Table 16. Even though some strategies have not been formally evaluated in the academic literature, these strategies may still be widely used in practice and considered to be effective by those implementing or using them.

Category 1: Direct or indirect approach to instigator (victim, bystander or managers)

Approaching an instigator of UB can be performed by a target, bystander or manager in the workplace; it can be done at the time at which UB occurs, or later. However, this strategy relies on a great deal of self-confidence and the ability to overcome fear of reprisal in those approaching the instigator. Strategies to improve confidence and ability to do this are discussed in Category 2, which focuses specifically on the act of approaching an instigator (either directly or indirectly).

Approaching an instigator as a victim or bystander directly and telling them how it felt when exposed to their behaviour would be an attempt at informal resolution. A direct approach can also take other forms, such as writing a letter to the instigator or approaching human resources to facilitate a meeting to attempt to resolve differences with the instigator. 124 Disciplinary strategies include a formal investigation of an instigator. This could include reporting an instigator to management and/or writing a formal complaint whereby organisations may investigate patterns of UB and deliver a warning, or other disciplinary strategies or action if required. 97 Often, this is according to legal and investigative requirements.

Disciplinary action, however, also requires receptivity from management to engage with the process. If some organisations do not have a culture that enables resolution of these issues, then individuals can seek help externally – a strategy discussed in Category 3.

Included sources mention informal direct approaches very frequently (see Table 16) but it is difficult to evaluate the success of these strategies as they tend to occur more naturally or as an outcome of other strategies. Only reporting and professional accountability interventions, such as those in Ethos or Vanderbilt, formally include direct approaches to instigators as a strategy and enable both informal resolution via a peer messenger (to encourage reflection) and behaviour change while including an escalation to formal investigation if poor behaviours persist. 97,98 However, other interventions that encourage speaking up or build confidence may also be indirectly attempting to encourage recipients of UB to engage in informal resolution.

Category 2: Improving confidence to come forward/speak up

Several strategies seek to improve an individual’s confidence – and thus ability – to come forward to report or challenge someone who is engaging in UB. Examples include assertiveness training, cognitive rehearsal and role-playing; these were all evaluated in the included intervention papers. One study conducted in Turkey sought to assess the impact of assertiveness training to tackle mobbing (defined as ‘unethical communication and antagonistic action directed by one or several people toward a single individual systematically’157). A rationale for including assertiveness training was the claim that when encountering mobbing, the majority of nurses ‘act passively’. 157 The authors found that, according to a self-developed mobbing scale, mobbing decreased after the intervention – decreasing from a score of 226.4±27.7 to 159.6 ± 47.9 – and assertiveness improved from 6.23 ± 15.6 to 17.0 ± 16.06 (measured by the Rathus Assertiveness Inventory). A second study combined assertiveness training with other strategies, such as education, and found an associated decrease in reported verbal abuse rates from 90% to 76%, as well as reduced feelings of powerlessness among nurses. 126

Cognitive rehearsal and role-playing involve practising scenarios to improve mental readiness for when UB is encountered in the workplace. 67,152 Cognitive rehearsal involves practising responses to UB, usually mentally, to change them from automatic towards deliberated, as a means by which to enhance coping and reduce UB escalation. Ten of the included studies used cognitive rehearsal, making it one of the most used strategies. Five studies54,63,65,148,149 attempted to replicate a study by Griffin (2004), undertaken with 26 newly registered nurses, drawing upon cognitive rehearsal as their main behaviour-change strategy. While 62 new nurses received the intervention, only 26 nurses provided outcome data. The intervention drew on both education and presentation of appropriate responses to 10 common UB scenarios – such as staff engaging in broken confidences, scapegoating, backstabbing, undermining and verbal affront. The strategy included the use of cueing cards covering appropriate responses, to which nurses could refer later. The researchers reported that 100% (n = 26) of nurses reported confronting instigators during focus groups that took place 12 months post intervention. The organisational retention rate was reported to have been positively affected but no comparison of pre- to post-retention rates was provided.

All sources drawing on cognitive rehearsal as a strategy reported positive results, except for a study by O’Connell et al.,54 which also drew on the strategy of role-play. Role-playing is similar to cognitive rehearsal but involves practising responses with other people, as opposed to purely mental rehearsal. O’Connell et al.’s54 study was conducted with military nurses in the USA and drew on Griffin’s work, incorporating education, cognitive rehearsal and role-play. 54 The NAQ was administered before and 3 months after their intervention and no statistically significant changes in experience of UB after the intervention were found. This lack of change, they stated, could have been due to senior and junior staff being mixed together, which inhibited engagement in role-playing due to the strict military hierarchy. 54 All other studies drawing on role-playing reported positive results. 107,137–139,141,142,156

One final evaluated confidence-improving strategy is keeping records, which is a strategy that individuals can use to enhance the likelihood they are believed when speaking up, as well as providing evidence about who was involved. This may take the form of notes, journals or screenshots if harassed electronically. One study recommended this as a strategy as a result of a workshop, stating that ‘session participants recommended instructing residents to document very clearly what was said or done, and by whom, in order to equip the institution with the necessary information to make an intervention’. 72 This study reported raising awareness of UB – but it is not clear if or how keeping records had an impact on subsequent behaviours.

Category 3: Improving ability to cope (victim, bystander)

Some strategies can be a means to improve an individual’s ability to withstand the negative impact of UB, a concept also referred to by some as resilience. 170

Category 4: Understanding prevalence of unprofessional behaviour (managers/leaders)

Many studies, prior to an intervention, assessed prevalence of UB in their sample as part of an evaluation. However, few studies drew on prevalence surveys as a strategy to identify where UB was occurring in order to better focus and tailor the intervention. 146,150,154,155 This strategy category involves understanding where and how UB is occurring, to allow contributors to be better targeted by other strategies. Three studies using prevalence surveys drew on the CREW intervention, which incorporated myriad systemic strategies for reducing UB;146,154,155 another non-CREW intervention study also drew on a bespoke and highly systemic approach that included action-planning, communication skills training and workplace rearrangement. 150 The CREW intervention draws on a custom-designed CREW survey called the ‘civility scale’, which assesses a range of civility-related metrics such as respect, cooperation, anti-discrimination and value differences. 146 As such, it does not test for prevalence of particular types of UB.

Use of this strategy only in more intensive interventions demonstrates that the use of surveys to understand prevalence is typically performed either before a larger culture-change effort to identify where ‘problem areas’ are or to identify baseline data against which to compare changes following an intervention. Use of qualitative investigation and interviews with staff may also enhance survey data by identifying specific contributors to UB present in organisations. One study drew on a comprehensive 4-month investigation of workplace culture at a UK ambulance service, using a mixed-methods approach as well as over 120 hours of one-to-one telephone interviews. 76 The study gathered sufficient data to understand how different professional and demographic groups experience UB in unique ways and identified many systemic factors propagating UB. This enabled the proposal of organisational and/or bureaucratic changes that could minimise aspects leading to frustration etc. in the workplace – in addition to individual-level interventions that are so often the focus. 76

Category 5: Improving teamwork (all)

We have collated strategies explicitly attempting to mitigate UB by increasing a sense of camaraderie or improving the ability to work together as those that try to improve teamwork. These include activities such as team-building exercises,141,142 communication training (which may lower the risk of conflict caused by improper or unprofessional communication64,134,150,154,156,164), problem-based learning139 and journal clubs or group writing strategies. 104,144 On team-building exercises, for example, Armstrong141 drew on a modified CREW intervention and included team-building exercises, facilitating discussions about workplace incivility and role-playing. Their examples of team-building exercises included ‘Anything Anytime’, which involved ‘providing a generic subject and discussing how it is viewed differently by different members of the group’, as well as ‘Geometry of Work Styles’, which ‘involves participants selecting from four geometric shapes that relate to a personality type’. 141 The authors found no differences in exposure to incivility post intervention but did find it was able to modify nurses’ self-assessed ability to respond to incivility. Overall, two out of three interventions drawing on team-building were reported to be successful, with the other reporting improvements to group cohesion. 142 One did not assess efficacy. 152

Five studies drew on the strategy of communication training. 64,150,154,156,164 All reported positive results except for Chipps and McRury,134 who reported training on ‘techniques for giving feedback’ and anecdotally reported an increase in a ‘learning community’ after the intervention; however, details were sparse. They found that reported negative acts nearly doubled post intervention, alongside a paradoxical decrease in self-identified engagement in bullying. A more well-defined intervention by Saxton64 involved collaborating with a certified Crucial Conversations trainer – who, over a 2-day intervention, drew on several techniques including videos, role-play and reflection to improve communication abilities. They found an increase in self-effectiveness immediately post and 4 weeks after the intervention and a self-reported increase in ability to address disruptive physician behaviour. 64

Problem-based learning involves identifying with and attempting to tackle real-life problems, often drawing on peer-to-peer teaching and learning. Problem-based learning was only used by one study, which happened to draw on live actors ‘acting out a situation in which one of the staff nurses was extremely uncivil to her two colleagues’ in which students ‘observed the scenario, provided written feedback on the effectiveness of the scenario, and participated in small-group discussions to debrief after observing the scenarios’. 139 Post intervention, the authors reported that participants felt better prepared to address incivility in the workplace.

Two studies drew on journal club/group writing techniques. 104,144 One sought to decrease negative workplace behaviours by enhancing voice and connection among nurses, with music-assisted meditation lasting 15 minutes followed by a 1 hour 45 minutes creative writing exercise. This took place once a week for 6 weeks (although the wording in the study is slightly unclear). The creative writing drew on several steps: ‘(1) an explanation of the writing group method, (2) an introduction to the exercise, (3) 20 minutes of writing, (4) an invitation to read their writing aloud and (5) the group listens and comments on what is liked and remembered’. 104 This study did not assess effectiveness but found participants reported it as a ‘positive experience’. The other study using this strategy involved 6-hour-long monthly meetings over 6 months and drew on article-reading as well as other activities in each session, such as ‘role-play’, ‘breathing techniques’ and ‘stress-ball activities’. The authors did not assess effectiveness. 144

Conflict-management training107,134,154,155 can overlap with communication training and focuses on communication styles, problem-solving methods and conflict-resolution methods that can de-escalate situations to avoid UB issues. 107,134 For example, one included intervention drew on teaching problem-solving methods that ‘transformed from competition to joint problem-solving, which is the hallmark of competent conflict communication’. 107 This study reportedly managed to improve conflict management and communication skills;107 however, another found that UB increased after the intervention but did not assess actual conflict-management abilities. 134 The two other studies using this strategy adopted conflict-management training on a flexible basis as part of the CREW intervention strategy repertoire.

Category 6: Social norm-setting (all)

The strategies in this category revolve around setting new expectations for behaviour in the workplace. Championing was used by five studies93,138,142,147,151 and involves encouraging certain individuals to espouse anti-UB values and behaviours to help drive changes in social norms and, sometimes (as in the case of Freedom to Speak Up Guardians) to act as trusted contacts to whom incidents of UB can be reported. 168 The impact of drawing upon champions is highly difficult to measure, as culture-change may be subtle and champions were always used alongside other strategies, often over a long period of time. For example, three of these studies used the Vanderbilt intervention approach, which draws on many strategies – including a reporting system for identifying problematic individuals. 93,147,151 In this approach, champions were identified from individuals already known to espouse the positive values being sought. 151 These individuals were then used as go-to people to help implement other intervention activities, train peer messengers and disseminate project messages to various staff types.

Other social norm-setting activities include role-modelling, which can be done by champions but is more typically performed by leaders seeking to act as a model for lower-level employees. 126,138 Positive role-modelling as an active choice is a strategy; however, negative role-modelling can be a contributor to UB. One study used workshops on role-modelling, delivered to nurse managers first in order to train them to be role models for other staff. 126 A second tier of workshops were then delivered to 20 regular staff to train them to be trainers to other staff and be role models for those who attend their sessions. All studies using role-modelling reported positive results as per their selected outcome measures. 97,126,127,138

Codes of conduct also seek to set expectations of behaviour in the workplace; seven included studies drew upon implementing new codes of conduct. 103,125,134,138,150,154,164 Codes of conduct often form the backbone of a wider intervention and are not often delivered on their own – rather, they set out policies and practices for the organisation. These can include what UB in the workplace looks like, how to tackle it, what resources are available to employees to address it and resolution pathways, among others. Sometimes, codes of conduct can be enhanced by increasing awareness of the policy or by asking employees to sign a pledge to abide by them during hiring or performance reviews. For example, one study drew on this approach, stating that ‘beginning in 2008, at the time of the annual performance review, each leader commits in writing to consistently holding staff, peers, and leader colleagues accountable to the Highland Promise Standards of behavior’. 138 This demonstrates that annual reviews can also be used as a time point at which to raise awareness of the code of conduct and to re-commit to it.

Category 7: Improving leadership competence and empathy (managers/leaders)

UB originating with management is unfortunately common, as statistics from the NHS staff survey and others indicate. 16 Within this category lie strategies that try to improve how leaders and managers interact with their employees and their management styles through leadership training. We did not include conflict-resolution strategies delivered only to management in this category – these are in the conflict-management category. Only one paper drew on some form of leadership or management training and this was using the Vanderbilt programme. 98 This paper drew on an academic leadership programme using role-playing techniques for skill development, as well as a programme called Elevate to promote professionalism in leadership. 98 However, this paper did not assess effectiveness. Despite a lack of management skills being a major cause of UB and UB originating with management being a core statistic within the NHS staff survey, for example, the dearth of papers evaluating this as a strategy was surprising.

Category 8: External pressure on organisations (managers/leaders)

External pressure on organisations can come in several forms, typically leading to managers or leaders gaining an increased sense of urgency to tackle the issue. Seeking hospital Magnet status for being a positive place to work can drive motivation for leadership when tackling UB and can indicate a path to achieve it. Two studies mentioned seeking Magnet status, which is a certification indicating a high-quality place to work – currently in use most frequently in the USA. 138,142 Achieving Magnet status requires exemplary performance in terms of transformational leadership, structural empowerment, professional practice, new knowledge generation and empirical outcomes;173 as such, reducing unprofessionalism is key to achieving this status. Magnet was the only such programme we identified in the literature, although other such ‘awards’ for hospital culture may also exist. Included interventions mentioned that Magnet was a motivating force for implementing interventions; however, Magnet was not an intervention in itself. Since the status is mostly a motivating force for managers and leaders, one critique of seeking an accreditation is that it may only change behaviour of managers and leaders. This is a feature of all the strategies that place external pressure on organisations within this category.

Category 9: Reporting and escalation systems (all)

Reporting systems involve implementing a process by which co-workers can document instances of UB (often online). These reports or messages are then reviewed by dedicated personnel to tackle the issues raised, usually in a graduated manner (which may then be escalated). Although it is possible to implement a more basic reporting system without a graduated component, we did not encounter this in the literature. Both the Ethos97 and Vanderbilt93,98,147,151 systems pair the reporting system with a means by which to escalate the approach taken to the instigator over time, typically starting with a message delivered informally by a peer messenger. If UB continues, then a more formal investigation process or management plan is initiated. The reported messages can be anonymous, which can increase willingness to speak up. 97 Vanderbilt has seen significant proliferation in the USA93,151 but included papers did not report on effectiveness of the intervention. One paper assessing Ethos found it to be associated with a reduction in the prevalence of UB. 127 We provide more details of Vanderbilt in the section below on the flexibility of interventions.

Category 10: Workplace redesign (all)

As mentioned in Chapter 3, when a workplace reduces job control and worker participation in decision-making, this can increase worker frustration, which is a contributing factor to UB. As such, redesigning the workplace can be considered to be addressing systemic contributors to UB. Despite this, only one included study sought to redesign work processes to facilitate worker involvement and increase job control. This study did not specifically refer to their strategy as democratisation – but in response to findings of a culture of intimidation, it did seek to include less senior employees more in decision-making processes around how to address this culture. 103 This, they reported, may have contributed to a reduction in staff turnover rate from 28.4% in the preceding 3 years to an average of 21.9% in the 3 years thereafter. Unfortunately, we were not able to find larger-scale workplace democratisation efforts in the UB literature in health care.

Category 11: Improving awareness and knowledge (all)

Education was the most common strategy employed by included interventions (see Table 16). It is also a strategy relatively easy to deliver, typically in a lecture-type format or ‘session’, or with use of simple leaflets. 65,72,145 However, the content of the educational sessions varied substantially. Most interventions provided educational content about what UB is, ways to cope or how to address it in the moment. Generally, information covered a wide range of UB types; however, some interventions mentioned including specific behaviours such as education about microaggressions. 72 Others drew on education about specific ways in which UB might arise; for example, one study drew on Structurational Divergence Theory (a theory suggesting UB arises due to inherently conflicting and opaque organisational structures)176 to teach participants that an underlying source of their workplace frustration might be a ‘treacherous institutional landscape’, which can sometimes be unnavigable. 107 This, they hoped, would change the way staff perceived UB and its origins; the authors reported reduced negative conflict attitudes and behaviours and greater empowerment to manage workplace conflicts among participants. 103 Most commonly, studies simply drew on education regarding UB and its effects, and some simple strategies for attempting to address it. 54,125 The use of education alone was not always successful in changing behaviour and, although such strategies often resulted in greater awareness of UB (which was perhaps the aim in many cases), this came with the unintended side effect of simply leading to more reports of UB. 66,134

Category 12: Implementation-aiding strategies (managers/leaders)

This category of strategies includes activities that may help sustain engagement or plan other behaviour-change strategies within a larger intervention, such as ongoing co-creation efforts. This can range from simple one-off exercises, such as where ‘full-day workshops were attended by more than 90% of nursing supervisors to receive feedback on the research and to develop compatible strategies for their areas’103 to large 6-month efforts in which action-planning is a constant feature, such as in the CREW programme. 154,155 The CREW programme offers a selection of activities from a repertoire that can be chosen for implementation depending on the situation. Indeed, development of such a repertoire of strategies or activities that can be implemented on demand can itself be considered an implementation-aiding strategy. While many of these strategies are formulated in advance and are part of the repertoire, others are developed as required for circumstances that may arise during implementation. 146 To facilitate this, meetings are scheduled to provide dedicated time every week for 6 months, to provide space and time for planning or implementing anti-incivility activities in addition to larger meetings at the beginning and the 3-month midpoint. This ensures there are no excuses relating to capacity for delivering programme activities. To facilitate these meetings, facilitators are chosen and trained to ensure that activities can be delivered in a high-quality and consistent manner by those passionate about tackling UB. The existence of dedicated time enables action-planning to occur, whereby the plans to address cases or contributors to UB are formulated. Many included interventions drew on action-planning. 93,98,103,134,137,154,164 Another implementation-aiding strategy can be the provision of dedicated personnel whose role it is to have dedicated time and space for helping deliver the programmes and initiatives. As such, all these activities serve to generate momentum. However, building in more implementation-aiding techniques is likely to require increasing organisational resource.

Category 13. Changing recruitment processes (all)

These strategies relate to changes that can be made to hiring, firing and recruitment processes, to try to remove or avoid hiring ‘bad apples’. As such, this is also a category of individual-focused efforts.

Key Dynamic 1: Interventions should address systemic factors that contribute to unprofessional behaviour, not only individual factors

As highlighted earlier in this report, many interventions ultimately assume that individual failings primarily cause UB. For example, interventions and associated strategies to improve awareness and knowledge assume that people do not know what UB is or how to address it. Likewise, strategies that improve confidence to come forward assume that people simply need to have more confidence to be able to speak up and address an incident of UB in the moment. This may be assumed to be easier to address than fostering a psychologically safe environment. As highlighted in Chapter 4, many contributors are organisational or systemic and are not likely to be solved with a solely individual focus. Our MRT of fundamental attribution error (FAE) (see Appendix 2 for more details) occurring in the minds of key leaders and intervention architects may explain how organisations inadvertently slip into the mindset of considering UB an individual-level issue while downplaying the role of systemic factors. 94 However, a focus on identifying and resolving individual attitudes and issues can, in many cases, lead to larger systemic issues remaining unaddressed. When these systemic factors are unaddressed, they can continue to lead people to behave poorly.

It is important to note that systemic issues are, of course, often more difficult to tackle. This is evidenced by the lack of workplace redesign efforts in included studies. One study highlighted a quote from a participant regarding efforts to reduce UB that their organisation was employing:

NGN1 explained: ‘I honestly think if we just had more staff and more resources, if we reduced the stress levels, that in itself would reduce incivility and bullying drastically. Because if you do that, people are less stressed and people are respecting each other more’. 177

This dynamic is highlighted in CMOC 39.

Key Dynamic 2: Focusing on individual staff can have unintended consequences for psychological safety

This dynamic builds upon dynamic 1 (above) and concentrates on the unintended consequences of focusing on individual staff members. Like Ethos or Vanderbilt, several programmes – while incorporating other strategies – have at their core the notion of ‘professional accountability’, that is, that it is an individual problem if staff are not behaving professionally. One reported side effect of interventions focused on reporting badly behaving individuals is that they create the conditions in which the default behaviour is to report colleagues when they are behaving badly, rather than seeking to resolve issues face-to-face as they happen. This also furthers the perception that the root cause of UB is problematic individuals, which can further absolve an organisation of responsibility or blame. The ability to report colleagues easily can inadvertently make it appear less necessary to tackle those conditions that make poor behaviour more likely (addressing other, more difficult-to-tackle, contributors).

One included study using the Vanderbilt approach highlighted such unexpected consequences, stating that ‘introducing feedback about co-worker concerns may have unintended consequences ranging from increases in unprofessional conduct and retaliatory reporting to negative effects on team cohesion and a culture that emphasises reporting over colleague-to-colleague conversations about concerns’. 151 The same study highlighted that recipients of negative messages generally did not take them well, demonstrating that it can be a somewhat traumatic experience to receive such reports: ‘recipients’ responses included blaming systems and other people, asserting inaccurate reporting, minimizing their behavior’s impact, expressing disbelief that three reports over three years constituted a pattern, focusing on who might have reported, and offering to “apologize”’. 151 Another study went further, to suggest that preventing an accused from resolving an issue with an accuser can be a form of marginalisation in itself:

an accused is ‘isolated’ from communicating about a complaint with other staff [...]. While a laudable goal, such an act can be experienced as kind of marginalisation, in itself a harmful practice which is understood by some to indicate management’s complicity with bullying acts. 47

Based on this excerpt, it is unlikely that this fosters a psychologically safe environment in which trust between colleagues is maximised (see CMOC 41). Indeed, one can argue that creating an anonymous reporting system simply allows psychological safety to be bypassed and, instead, seeks to maximise ease of ‘speaking up’ (Figure 18).

FIGURE 18.

Locating bad apples is not the same as enhancing psychological safety.

It is important to note that the literature suggests that identifying problematic individuals may be very effective at reducing incidence of UB, despite the risk that such systems may also be used to undermine others. Individuals may have lesser or greater proclivity to engage in UB due to their personalities or susceptibility to systemic issues such as pressure or lack of job resources etc. As such, reporting systems may eventually ‘weed out’ people from the environment; this may have the effect of changing the culture towards one in which UB is not tolerated.

Key Dynamic 3: How and why an intervention is expected to work must be clear; otherwise, evaluations of interventions can be misleading

Many of the included studies used intermediate outcome indicators – such as ‘assertiveness’ or ‘communication quality’ – as a measure of overall success. However, these are not sufficient to determine whether UB has been addressed. This issue may partially be due to underuse of theory and logic models when designing these interventions. One pertinent example of this is use of ‘awareness’ to assess intervention success.

Reliance on measures of awareness or prevalence of UB to gauge the success of an intervention is problematic. For example, four included interventions that relied on reports of UB as their primary outcome measure were reported by their authors as being ‘unsuccessful’. This was due to an increase in reports of UB seen in the intervention groups post intervention. 66,134,153,158 As the authors state, an increased number of reports of UB means that awareness was increased, which these authors consider in the ‘limitations’ sections of their studies. These authors also consider it to mean that their interventions are unsuccessful; however, this may be an erroneous judgement. If interventions were more robustly designed with a logic model that took intermediary concepts such as ‘awareness of UB’ into consideration, then improved awareness that led to staff increasing their reports of UB could be considered a successful example of behaviour change. This could then lead to more incidences of UB being addressed in the moment.

It is also important to note that while there are strategies that specifically seek to raise awareness of UB, many strategies do not. However, many strategies that do not aim specifically to raise awareness of UB are likely to do so as the intervention will raise the presence of UB as an issue in participants’ minds. As such, relying on decreased awareness or reports of UB to indicate intervention success may be problematic. To properly assess intervention effectiveness, the use of a wide range of outcome measures is required that enables an understanding of how and why awareness is key to reducing UB and improving patient safety (Figure 19). CMOC 42 highlights the risk intervention architects may encounter if they do not rely on robust logic models.

FIGURE 19.

Figure demonstrating that much of how interventions are intended to benefit UB or patient safety is left unclear.

Key Dynamic 4: Maintaining a focus on why it is important to reduce unprofessional behaviour (e.g. to improve patient safety) is key when designing an intervention to reduce unprofessional behaviour

We identified a dynamic whereby interventions may seek to reduce UB for the sake of reducing UB – but do so without enhancing ability to speak up in the moment and/or without improving staff conditions or patient safety. While reducing UB is the main goal of such interventions, it is important to remember that the primary reason to reduce UB should be to improve the psychological well-being of staff, and improve patient safety and quality of care. Improving ability to speak up in the moment can often be essential to improving patient safety14,24 (Figure 20); thus, interventions should seek to improve psychological safety. Included studies highlighted in some scenarios that, despite a programme being implemented to reduce UB, managers were still dismissive of patient safety concerns: for example

FIGURE 20.

Highlighting the juxtaposition between enabling speaking up in a way that enhances patient safety vs. a way that is less impactful to patient safety.

On multiple occasions over the past six months, [managers] have been disrespectful and dismissive of staff when patient management and safety issues have been raised …. Whilst I applaud the new safety system … the cultural change looks like it will take much longer. 93

However, such concerns should be an impetus for attempting to reduce UB in the first place.

Since many interventions discussed in this report involve rather distal aspects to improving patient safety – such as trying to increase staff assertiveness, for example – it is theoretically possible that novel methods such as CBM (discussed in full in Appendix 7) may have a greater direct impact on patient safety. CBM might be a valuable addition as it focuses on improving information-sharing, which is essential to maintaining safe patient care. 14

Included studies were not able to demonstrate that reducing UB improved patient safety or staff psychological well-being and that doing so is quite difficult since patient safety and well-being are a result of many factors. Despite the recognised importance of a patient safety basis for reducing UB, studies rarely attempt to measure improvements on patient safety metrics. Likewise, evidence that attempting to improve the ability for staff to cope with UB has an impact on the organisational culture or workplace design is absent. Very few interventions attempted to measure worker psychological well-being, while some drew on turnover rate as a proxy for worker satisfaction. While turnover rate may give some indication of staff satisfaction with their workplace, it is likely to also be a result of myriad other factors.

Key Dynamic 5: Encouraging bystanders to intervene is important for culture change but can lead to moral injury

Encouraging bystander intervention was a common strategy in the direct approach to instigator category in Chapter 6. Encouraging bystanders to intervene can send an important message that UB is no longer tolerated in an organisation and can lead to UB being addressed in the moment. However, encouraging bystanders to intervene can create a sense of moral imperative and may lead to moral injury (a MRT we outline in Chapter 3 and Appendix 2) if bystanders do not then have the courage or skills to intervene the moment UB is occurring. Alternatively, they may intervene and experience reprisal – or inadvertently escalate the conflict. A debate between demanding bystander intervention and being wary of it was frequently present in the literature. On the one hand, sources stated ‘We believe health professionals have a moral duty to practice ‘upstanding’ — intervening as bystanders’178 and on the other, ‘We thus find some of the advice offered rather out of place, e.g. for a bystander to confront a bully, given that the main remit of any bullying intervention should be to encourage harmony, not discord, in a workplace’. 47 This debate is reflected in CMOCs 44 and 45.

Key Dynamic 6: Identifying unintended consequences of anonymous reporting systems is essential

While allowing anonymous submissions about UB may increase speaking up behaviour, it may also increase the chance that systems may be misused. Systems that enable other people to be easily reported – especially anonymously – sidestep direct person-to-person resolution and may be used for purposes of undermining and scapegoating (CMOC 46). Sources highlighted the potential misuse of reporting systems: for example, one reporting system used within an adapted Vanderbilt approach identified how ‘a few staff described feeling distressed after receiving weCare feedback based on what they deemed vexatious and fictitious content’. 93 Another example arose from the Ethos intervention, which itself was ‘reported’ to the media for, among other issues, fostering alleged ‘implicit sexism’. In this case, ‘Ethos data from St Vincent’s public hospital for the 12 months to this month, seen by The Age, shows female medical staff were the subject of about 30 per cent more complaints than male medical staff’. However, it is also possible that going to the media regarding the intervention was itself an attempt by entrenched cliques to avoid a problematic culture from being addressed. Undermining of intervention measures seeking to address UB by cliques is a common theme, as this example demonstrates:

The continued emphasis upon in-house reporting schemes as a central plank to address workplace bullying inadvertently risks the protection of perpetrators, as it is possible that actors in alliances m0ay work together to subvert organizational reporting processes to protect their own interests. 117

Similar unintended consequences have been identified with the UK Datix electronic incident-reporting system; an incident-reporting system intended to aid patient safety used within NHS organisations. 179 Although primarily focused on patient safety, Datix is also known to be used to report UB between healthcare staff. For example, a recent publication highlights that ‘several high-risk events, including the assault of my team member by another staff member within a forensic unit, mandated the completion of an online Datix form’. 180 Informal accounts via social media provide some examples in which staff (Figure 21) report the weaponisation of the Datix system to the point where ‘Datixing’ has become a verb and is used as a threat. 180

FIGURE 21.

Tweet depicting the Datix system being used intentionally or unintentionally to undermine a doctor in practice in the NHS.

Bal et al. 180 highlight the weaponisation of Datix, stating that

45% of reports attributed blame to an individual, potentially reflecting an organizational culture in health which leads to retribution, rather than one of learning, along with ‘a failure to appreciate the contribution of system factors’, undermining the original purpose of the incident reporting system. 180

These dynamics are reflected in CMOCs 46 and 47.

Key Dynamic 7: Interventions must be perceived as authentic to foster trust in management

Many of the key dynamics we identified relate to building trust in management, which is often undermined when staff see UB occurring around them, unaddressed.

Staff make a constant assessment of the culture of an organisation, particularly in terms of UB. Staff will notice behaviours by managers and take these as a signal for what is tolerated and acceptable. In organisations that have historically had an issue with UB, it is likely that staff strongly desire change towards a civil environment and seek signs that this is the direction of travel. This ties in with trust in management (a MRT outlined in Chapter 3 and Appendix 2), because staff who experience unaddressed UB for a prolonged period of time will lose trust in management. 48,143,155 To assess whether it is worth trusting management to provide a safe working environment, healthcare staff will assess the authenticity of efforts that management make to reduce UB. By authenticity (and genuineness) of the intervention in this section, we mean the perception of ‘whether there is a collective change in management intention to properly address UB and regain employee trust’. This notion of a ‘changing wind’ was highlighted by another included study: ‘The commissioning of the interview study and the wide public sharing of the findings functioned as an intervention in its own right, demonstrating to sceptical colleagues a new appetite for learning in the organization’. 143

An assessment of authenticity may involve looking at the intervention itself and assessing whether its components are likely to have the desired effect (i.e. is it a ‘tick box’ or a real attempt at changing culture?) (Figure 22). For example, one included study found that

FIGURE 22.

Perception of programme intervention authenticity.

employees’ attributions and perceptions of the organisation’s intention to prevent negative employee behaviours (bullying) only have their desired effect when policies are perceived to be implemented effectively. Effective implementation appears to signal to employees that bullying behaviours will not be regarded as acceptable behaviours within the organisation. 181

Staff will also look to managers to ensure they are not continuing to role-model negative behaviours themselves. One included intervention was undermined in this way: ‘at mid-intervention several participants had observed leaders behaving in an unprofessional manner which was sending inconsistent messages and perceived as impeding implementation’. 93 Likewise, another study reported interventions to be less successful if an organisation was known to be poor at following up and sustaining initiatives. 21 As such, the sense that interventions are genuine must be fostered, otherwise healthcare staff will not properly engage with interventions because they will not trust management and not trust that it is worth their time. This will undermine the intervention’s effectiveness.

Key Dynamic 8: One size does not fit all – tackling unprofessional behaviour generally requires multiple and sustained interventions to address underlying contributors

Related to Key Dynamic 7, we also identified that doing something to tackle UB is not the same as doing enough. Interventions will likely require many strategies, sustained over a long period of time and with genuine systemic changes to address contributors leading to UB. One study, for example, identified that their one-off educational intervention was insufficient to address UB in their organisation, stating ‘a one size fits all approach is not sufficient to mitigate negative workplace behaviour and that using education as a “tick box approach” is unlikely to be effective in developing sustainable cultural changes in organizations’. 177 This, however, relies on sufficient resource being available to implement follow-up action that draws on more strategies. Organisations also run the risk of becoming complacent because they are already doing ‘something’ to address UB, which can lead to other contributors remaining unaddressed. In these cases, for example, ineffective policies or procedures are seen as adequate and can lead to situations in which ‘workplaces with prominently posted policies that mandate respect or decry bullying often fail to address actual complaints’. 114 One study also demonstrated this, stating that a bullying intervention was ‘management’s “one hit wonder”’ but that it ‘would require concurrent and similarly successful initiatives to raise their level of trust in management’. 122 This dynamic is highlighted in CMOC 50.

It is also important to note that failing to tackle systemic issues, such as unfairness in the workplace, can disproportionately impact those individuals with protected characteristics who suffer from systemic or less visible forms of UB, such as discrimination. One included study highlighted this, stating that

race equality and inclusion may be unamenable to simple applications to evidence-based interventions that do not take into account the subtleties of local context and history. A ‘do-once’ attitude, or well-intended processes that do not model the spirit of the outcome, may even be harmful. 131

Key Dynamic 9: Addressing manager behaviour is essential for building trust in management

Interventions work best when they are seen to include management; it is essential for trust in management to be (re)built (Figure 23 and CMOC 51). If an intervention involves ‘targeting’ individuals – such as with a reporting system – then it must include the ability for managers to be targeted, too. If it delivers training, this is less important but still helpful when it comes to demonstrating that management is involved in moving the culture in a positive direction in some way. The importance of this is explained in one study:

FIGURE 23.

Managers must implement programme interventions that include themselves.

Management participation is also important because these staff are not immune to bullying behaviours, and in some cases are central protagonists. Either way, ‘what management do’ (role modelling and showing they are learning and supporting other staff) has been shown to vastly influence workplace culture and any intervention. 22

However, managers or leaders are usually the ones who implement any intervention to reduce UB (CMOC 52). Thus, from the outset, there’s already a potential incentive to exclude themselves being targeted or involved in the intervention, creating a dilemma that organisations need to address. This will be especially important for those organisations in which managers have been seen to engage or tolerate UB themselves, with subsequent low levels of trust in management. In cases in which managers do not want to be targeted by the interventions they implement, interventions may not be successful; in such cases, the only resolution may be external pressure from regulatory bodies or enforced leadership changes.

Key Dynamic 10: Interventions that are both inclusive and equitable are critical to ensure effectiveness and sustainability and for addressing inequalities

Interventions should be seen as both inclusive and equitable. However, we identified that it was rarely possible to achieve both, particularly when it came to creating equitable solutions across different groups. As discussed earlier in this report, certain groups with protected characteristics experience more UB in the workplace. Yet these groups are currently rarely catered for in existing interventions to tackle UB. One could argue that this reduces fairness and equity of access. For example, the following excerpt from one study notes: ‘despite their selflessness and arduous work, Black African nurses face structural and institutionalised discrimination within the NHS. Employers must challenge the dominance and hegemony that exists within the NHS to ensure greater equality of all employees’. 69 Interventions could – and arguably should – be more accessible and designed and specifically focused on reducing UB for these groups. One way to do this could be to deliver an intervention to improve opportunities for these staff with protected characteristics or to deliver an intervention specific to racism.

However, at the same time, included sources have highlighted that being inclusive is also important. While equity is essential to the success of interventions, it is also important to include as many people as possible in an intervention and not target one group over another. This is because targeting interventions at specific groups could alienate certain groups or imply they are at fault. Examples could include targeting doctors with an anti-bullying intervention because they are perceived to be bullies more often than nurses. One study highlighted that

it is now noted that some methods of delivering interventions to staff may induce feelings of being ‘targeted’, ‘at fault’ and perhaps being bullied themselves, if content is ‘aimed’ at certain negative behaviours, say ‘anger management’, or staff groups, say ‘the doctors’. 47

This highlights that this dynamic can apply to professional – as well as demographic – groups (e.g. gender).

One included study highlighted a participant reacting to the existence of a leadership programme open only to staff from an ethnic minority background:

For the [ethnic minority] leadership programme, it’s well, ‘What [have you] been taught which I couldn’t be taught?’ What is actually being said [to programme participants]? Is it, this sounds really bad and really racist, but is it, ‘Don’t listen to the white man’?. 131

In addition to the perception that such interventions are unfair, sources also reported that deficit-led interventions may have unfortunate connotations as ‘development programmes aimed at ethnic minority staff might create a sense that they somehow need to be “brought up” to the same level or standard as white people’131 (see Stakeholder feedback summary – inclusion and equity). As such, these dynamics highlighted a tension between inclusion and fairness, and that picking one or the other may result in negative consequences (Figure 24). These dynamics are also reflected in CMOCs 53 and 54.

FIGURE 24.

Interventions that are both inclusive and fair are difficult to achieve.

A further risk is that delivering interventions for only one marginalised group may inadvertently also exclude those from other disadvantaged backgrounds, such as those from an economically deprived background, LGBTQ+ people or people with a disability. It also shows that interventions to enhance fairness at the cost of inclusion may generate unnecessary conflict. Simply ensuring all employees have equal access to opportunities (equity) may be the less risky option for organisations. It is not clear whether a lack of interventions seeking to address UB towards marginalised groups is due to leaders wrestling with this key dynamic or simply due to underinvestment and lack of prioritisation of this issue.

Key Dynamic 11: There are trade-offs between fixed interventions and flexibility

As discussed in Chapter 5, we identified that some interventions were adaptable to the situation and organisational context. This meant that certain strategies might be used in some contexts but not in others. Theoretically, this adaptability can enable enhanced efficacy, because drawing on a repertoire of strategies can mean strategies can be chosen that better fit the context (CMOC 55). For example, Laschinger et al. 155 set out how certain strategies can be used to tackle certain goals in the CREW intervention. In CREW, weekly meetings enable identification of the challenges ahead and strategies can be either drawn from a repertoire or created on demand, to be implemented on a case-by-case basis. An example they provide is one in which junior managers identified that a lack of conflict-management skills in their unit may have been leading to UB. As such, discussion in a weekly meeting led to implementation of role-playing scenarios to build the missing conflict-management skills. 155

This trade-off between flexibility and fidelity necessarily means that it is more complex and difficult to be able to assess why the intervention worked since the intervention will not be the same when delivered in different contexts. For example, was success due to implementing this conflict-management strategy or was it due to another strategy targeting another situation – or did it work despite implementing these strategies? As such, when implementing CREW in one organisation versus another, the actual implementation may comprise entirely different delivered components and thus be similar to comparing two totally different interventions in behavioural terms. This difficulty of evaluation is reflected in CMOC 56.

Key Dynamic 12: There are trade-offs between a theory-first and practice-first intervention design

In Chapter 5, we identified that many interventions lack a strong theoretical framework and logic model, and that many interventions are being conducted in practice and are not reported in the academic literature. Key Dynamic 12 refers to two main ways we identified to design an intervention: (1) top-down (i.e. drawing on robust theory, logic models, modelling the assumptions and components behind how UB may occur, developing an intervention that targets the antecedents of the behaviours) or (2) bottom-up (i.e. identifying a problem on the ground, searching for an applicable solution and implementing it to prevent UB as quickly as possible). Academics may use the former top-down strategy, while many working in practice in the NHS often use the latter strategy.

A top-down approach may have more ability to test what works and why, whereas the bottom-up approach may provide adequate results more rapidly. One limitation of a top-down approach could be that it does not properly target the contributors ‘on the ground’, as highlighted by one study:

The Program’s implementation was commonly viewed as being ‘top-down’ and not adequately reflecting the unique needs of unit-level sub-cultures: ‘I feel that the staff aren’t really connected with the organisational programme because it kind of sits up here and it’s very disconnected from what goes on at grassroots level’. 93

While we are portraying this as a key dynamic, we do believe it is possible for a theory-based, well-articulated and practical intervention to be constructed. However, this will require suitable guidance to be in place that would enable practitioners on the ground to easily access information required for tackling UB in the workplace (see draft manager’s guide, Appendix 8). This includes which strategies target which contributors and a basic introduction to changing behaviour, as well as factors that can inadvertently affect intervention implementation such as these key dynamics. CMOCs 57 and 58 highlight the key trade-offs.

Implementation Principle 1: Reach

It is important that interventions seeking to address UB have broad reach. This means including all staff groups, including management and senior leaders, to demonstrate a serious effort by the organisation. 151 Evidence suggests that including as wide a range of staff as possible could also provide additional benefits, such as improving interprofessional interactions22 and reducing ostracisation of certain groups, as discussed in Key Dynamic 9. However, consideration must be given to how different groups are brought together to promote psychological safety; particularly if hierarchies are mixed (see Implementation Principle 15).

Implementation Principle 2: Co-creation

Including staff in the design and delivery of intervention content can be a practical way to increase a sense of ownership among staff. 119 Engaging staff in this way can also help to empower them (by making them feel involved in doing something to tackle the issue),138 as well as ensuring that intervention content is tailored appropriately to ensure uptake. 138 Examples in the literature indicated that staff can both help plan the intervention and become agents of change during intervention implementation. 126,138 Similarly, co-creation can help increase a sense that managers are now serious about addressing UB and that they are making an authentic effort to design a proper intervention. This can help to resolve Key Dynamic 5 and can help sustain engagement with the intervention for longer.

Implementation Principle 3: Assess before implementation

Identifying organisational antecedents is an important initial step. Surveying organisational culture and identifying the contributors to UB within an organisation can enable better tailoring of strategies to an organisational context. Assessing organisational context before implementing any interventions or strategies can be useful in multiple ways: it can identify pockets of UB that may require more targeted approaches, enable an understanding of how the existing culture is contributing to UB, and identify specific UB that require specific tailored interventions or strategies. In the CREW intervention, for example, the pre survey determines how a site can best implement the intervention. Conversely, one included study highlighted that ‘failure to identify and address catalysts could itself be understood as a way to “facilitate” or perpetuate bullying’. 22

Implementation Principle 4: Dedicated staff to lead work to tackle unprofessional behaviour

Identifying dedicated staff to tackle UB can facilitate intervention sustainability and provide a single point of contact for organisational expertise and knowledge, thereby enhancing its success. One study referred to drawing on an ‘interventionist’ to ‘develop a knowledge and understanding of what exactly staff do and a department’s clinical function within the health system’ to ‘offer appropriate, applicable content’ for an intervention, as well as to ‘ensure the offering is accessible’. 22 Use of dedicated staff can increase the perception that management is engaging in a genuine effort to address UB (Key Dynamic 5).

Implementation Principle 5: Skilled facilitation

Many types of intervention require a facilitator to deliver content. This can include the use of training strategies, such as conflict-management training, educational activities or the training of essential staff members as mentioned above (e.g. peer messengers97). Comprehensive training enables foresight of any issues that may crop up during intervention delivery, avoiding unintended consequences before they happen.

Implementation Principle 6: Multiple simultaneous strategies

Use of many simultaneous strategies (as opposed to only one or two) in an intervention was often cited as a major factor in success. 68,93,138,141 Purported reasons indicate that use of many strategies ‘link together to form social systems that promote changes in behaviour norms’172 and that strategies ‘must be multidimensional in order to reach all constituents in an organization’. 138 It is thought that the use of multiple strategies may be more effective at changing culture and practice. 68,141 This works best when there is a full understanding of both what strategies are appropriate for the type of UB (see Implementation Principle 3) and why and how those strategies may interact and complement each other. Evidence for this is currently lacking. Additionally, inclusion of a greater number of strategies will incur a resource cost. Nonetheless, drawing on multiple strategies can help to improve intervention authenticity (Key Dynamic 5) and address a greater number of contributors to UB (Key Dynamic 6).

Implementation Principle 7: Maximise visibility

Interventions must be highly visible. Maximising visibility of activities or interventions to address UB sends a strong signal about leadership and culture change and can improve effectiveness. During the project lifecycle, advertising successes of the intervention can also be extremely important: for example, in organisation-wide e-mails from leadership or organisational newsletters. This can help sustain momentum and build trust in the intervention, which can help build a sense of authenticity essential to Key Dynamic 5. For example: ‘The Programme is well advertised, information about it is everywhere – posters, intranet page, computer screen savers. I think this helps provide a strong stance that this programme and changing our culture is a priority’. 93

Implementation Principle 8: Early intervention

Intervening early – as soon as any negative patterns of behaviour are identified – was often recommended as the best approach to send a signal that UB is not tolerated. Yet, implementing an extremely well-designed intervention, drawing on many of the principles laid out in this section, can also be in competition with intervening quickly (see Key Dynamic 12). Nonetheless, intervening quickly is important because any delay can unintentionally give the impression that management tacitly tolerates or has a hands-off approach towards UB. Additionally, any potential impact on patient safety and care quality must also be addressed rapidly. It may be possible to implement interim measures in an organisation while a more robust and complex intervention is developed.

Implementation Principle 9: Manager and leader engagement

The engagement of managers and leaders with the intervention is essential both to ensure the content of the intervention is properly delivered and to ensure staff see that reducing UB in the organisation is a priority of senior staff. This relates to Key Dynamic 5 (enhancing authenticity) and Key Dynamic 7 (engaging managers). For example:

… If senior personnel … promote and encourage [a programme to address UB], that will certainly have a positive impact … and vice versa, if there’s not a lot of interest shown … then people under them are less likely to be engaged and see it as a valued program. 93

Implementation Principle 10: Intervention perceived as just and not punitive

It is important that interventions are perceived to be justified and not overly punitive or staff may become apathetic and disengage from the intervention and its content. For example, ‘Just culture is also supported by sharing and addressing behaviors that undermine teamwork by promoting collaboration, accountability, self-evaluation, and decency to help solve system issues and create a blame-free error reporting system that supports peer feedback without punitive measures’. 53 It is also important to avoid an intervention undermining psychological safety, which links to Key Dynamic 2. Natural justice – the notion of ‘innocent until proven guilty’ and the ‘right of reply’ – is also essential in specific strategy types, such as with reporting systems. In these, there are graduated systems depending on whether there is a one-off negative behaviour or a pattern of UB. It is important that the perception of justice in the system is maintained to ensure that one-off or less severe events are not treated the same way as patterns of behaviour or severe forms of UB. For example, ‘anonymity has led to frivolous complaints and some staff feel a divide growing between the messengers and ordinary staff. Ethos is providing a system of no accountability for complainants, no just hearing and no records for due process’. 182

Implementation Principle 11: Maximising existing opportunities

Existing organisational processes such as onboarding processes, appraisals or development reviews can be used to reduce UB. Although we did not identify any interventions doing so in the literature, some sources did mention the possibility of using existing organisational processes. For example, ‘the use of a pre-existing educational conference block to ensure protected time for fellows was helpful in seamlessly integrating the curriculum into a busy academic fellowship programme and division’. 71 Our stakeholder group also highlighted the importance of tying in with ongoing efforts that are related to addressing UB in organisations, such as the Freedom To Speak Up Guardians. 168 Since these events and activities are already taking place, it can be low cost and low risk to use them to help drive behaviour change. Included studies cited drawing on onboarding processes, for example, to introduce the signing of ‘conduct pledges’, which can ensure awareness of the behavioural policy and make it easier to discipline those who behave badly in the future.

Implementation Principle 12: Organisational turnover and change

Staff and leaders who were championing an intervention and then leave an organisation can take much experience and momentum with them. The same can be said for processes of organisational change, whether that is changes in leadership or organisational mergers, which can serve as a distraction from efforts to reduce UB. It is important for those implementing UB interventions to plan for succession, intervention continuation and sustainability in cases of disruption to processes or staff.

Implementation Principle 13: Tackle the instigator of unprofessional behaviour

Literature reported that in cases where conflict was occurring between two people, some organisations simply moved the instigator or the target of UB to another department. This enables the instigator to continue his or her behaviour, either in the same or in another setting, and sends the signal that the organisation is tolerant of UB. This was cited in the literature as occurring most frequently when an instigator is a high-performing doctor or surgeon.

Implementation Principle 14: Incorporate ongoing evaluation

Incorporating ongoing evaluation into an intervention may increase resource costs but it can also enable greater adaptability that can enhance efficacy either while the intervention is running or when implementing new versions of an intervention (Plan, Do, Study, Act cycles). It can potentially save costs in the longer term because an evaluation can identify contextual factors that may have undermined its success. An evaluation can help with understanding the setting in which an intervention is implemented and improving understanding of contexts that influence its success, as well as proximal (i.e. psychological safety), intermediate (i.e. reports of UB) and distal outcomes (i.e. patient safety metrics) (Key Dynamic 11). Evaluation results can also demonstrate to staff that interventions are effective and thereby enhance both trust in the organisation and support for the intervention.

Implementation Principle 15: Do not mix hierarchies

While managers should be involved in interventions, it is important to also highlight that hierarchies have a large impact on ability to speak up (as outlined in Chapter 4, Category 1). As such, mixing of hierarchies can inhibit psychological safety in group-based interventions, undermining the ability of lower-level staff to engage with the intervention. This is purportedly a reason why one included intervention did not meet its objectives (see quotation, below). This principle is reflected in CMOC 73.

This principle could be resolved by designing an intervention that is seen to include managers in another way without mixing hierarchies, thereby avoiding this pitfall.

Table 19 provides an overview of our key implementation principles, mapped on to the consolidated framework for implementation research (CFIR), to facilitate stronger links to implementation theory.

Terminology

As highlighted in Chapter 3, terminology used to refer to UB is highly inconsistent and can present issues in practice, as well as when attempting to synthesise literature in this area. Definitions should be provided when seeking to assess the incidence of certain forms of UB or when referring to UB in a study. The academic and practitioner space would greatly benefit from a more uniform and clear understanding of the range of UB terms to help avoid such issues, as would staff in NHS organisations seeking to address UB. Drawing upon an understanding of the differences in dimensions of UB types may help shed some light on essential differences between them. To this end, we hope our typology outlining the dimensions and breadth of UB-related terms in the healthcare space will be an important first step towards the goal of mapping UB terms. Although it is not possible or advisable to dictate how UB should be defined, future studies should work on compiling a compendium of terms: perhaps through use of the Delphi method, for example. We believe our initial definition or understanding of UB, presented in Chapter 3 and refined with our stakeholder group, will help serve as a basis for future efforts in this area.

Behavioural science

More robust logic models (and indeed, programme theories) may also include depiction of how various intervention strategies are anticipated to interact and may even pre-empt implementation challenges and assumptions, and how to overcome them. No studies drew on modern theories originating in behaviour-change psychology, such as the COM-B approach (capability, opportunity and motivation for changing behaviour)190 or implementation science frameworks or theories [e.g. CFIR, i-PARIHS (Integrated Promoting Action on Research Implementation in Health Services),191,192 normalisation process theory193]. Future studies must fully explicate how and why they think the intervention components will lead to the outcomes they intend, including how implementation challenges in diverse contexts will be addressed. Future research may need to draw on multiple theories to explain how and why their intervention is intended to drive the desired effects. As our review indicates, mechanisms and outcomes are often overlapping.

Enhancing ability to synthesise findings

In addition to not adequately explaining how and why interventions are intended to work, reports of evaluations of interventions should give greater priority to reporting implementation context and how it could impact effectiveness. This will enable future evidence syntheses to gain a greater understanding of why a particular strategy may work in one context but not another.

Need for tools to assess UB prior to intervention

For interventions to address the actual contributors to UB, they need to first understand what they are. Our review identified that this was a limitation with existing interventions, yet few tools and instruments exist that are adequate for determining anything more specific than the prevalence of UB prior to intervening. 146 Tools should be formulated that provide greater insight into what is contributing to UB in an organisation and where it is taking place. Such tools may need to be mixed methods in nature, to capture both the prevalence quantitatively and ‘how or why’ it is occurring qualitatively.

Other healthcare settings

While our research sought to identify work around UB in acute care, we believe our findings should be contrasted and understood against other healthcare settings, such as private health care or primary care, too. This will enhance applicability of our findings while supporting research in other areas.

Reactive strategies