Remember me
REVIEW ARTICLE
Year : 2022 | Volume
: 26
| Issue : 4 | Page : 207-224
Identifying predictors of workplace violence against healthcare professionals: A systematic review
Archana Kumari1, Piyush Ranjan2, Siddharth Sarkar3, Sakshi Chopra4, Tanveer Kaur1, Upendra Baitha1
1 Department of Obstetrics and Gynaecology, Treatment Centre, All India Institute of Medical Sciences, New Delhi, India
2 Department of Medicine, Treatment Centre, All India Institute of Medical Sciences, New Delhi, India
3 Department of Psychiatry and National Drug Dependence Treatment Centre, Treatment Centre, All India Institute of Medical Sciences, New Delhi, India
4 Department of Home Science, University of Delhi, New Delhi, India
Correspondence Address:
Dr. Piyush Ranjan
Department of Medicine, All India Institute of Medical Sciences, Ansari Nagar, New Delhi - 110 029
India
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/ijoem.ijoem_164_21
Understanding the predictors of workplace violence amongst healthcare professionals is important to develop and implement prevention and mitigation strategies. We conducted a systematic review to synthesize the recent evidence on predictors of workplace violence across healthcare settings. The review has been done as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Two electronic databases (PubMed and Google Scholar) were used to search peer-reviewed studies published for the year 2009-2020 to identify studies reporting predictors of workplace violence. The significant predictors were analyzed using descriptive statistics such as proportions in most of the studies and some studies used inferential statistics such as logistic regression analysis, Chi-square test, ANOVA and Student's t-test. A total of 46 studies were identified and overall evidence was graded using an adapted GRADE approach. Some of the moderate quality predictors associated with workplace violence were the patient with a history of mental health disease, psychiatric setting, professional's gender and work experience and evening shift workers. Being a nurse was the only high-quality predictor. Healthcare professionals and administration can identify the predictors relevant to their setting to mitigate episodes of violence against healthcare personnel.
Keywords: Healthcare personnel, predictors, systematic review, workplace violence
Workplace violence (WPV) against healthcare professionals is an occupational hazard. Globally, at least one in five health care professionals experience physical violence annually.[1] According to the World Health Organization (WHO) WPV are “incidents where staff is abused, threatened, or assaulted in circumstances related to their work”. These violent episodes have direct or indirect impact on the employees' safety, health, and overall well-being. The impact is not just restricted to physical injuries, but extends to psychological well-being, which ultimately limits their work performance, job satisfaction, and patient care-related outcome measures.[2] Considering the impact of violence on professionals, it is important to identify the factors that compound to these episodes to devise integrated prevention and management strategies applicable in everyday clinical practice. Previous studies report that an episode of violence is an interplay of a number of factors such as characteristics of healthcare professional and perpetrator, setting, inefficient service systems, communication and interpersonal skills of the healthcare provider, societal norms, etc.[3] A few systematic reviews have focused on assessing the predictors of WPV with limited focus on specific aspects such as healthcare professionals (doctors or nurses),[4] departments (emergency),[5] and type of factor (organizational).[6] Even though a large number of significant predictors have been identified by these reviews, these predictors might not be applicable to all the healthcare professionals working in a hospital setting. In literature, a consistent set of predictors associated with violent episodes against healthcare professionals is lacking. We have taken up this review with the aim to identify the predictors of WPV in the healthcare setting.
MethodologyThe review has been done as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyzes (PRISMA) guidelines.[7]
Literature Review: In this systematic review, WPV was considered as any construct that was indicated as a violent episode in which the staff was threatened, abused, or assaulted in work-related situations, which included travel to and from work, involving a direct or indirect challenge to their safety, well-being, or health.
Search terms were identified through literature review, in-depth discussion and expert opinions. The authors devised a search strategy combining keywords pertaining to violence, healthcare professionals and predictors in the form: (aggression OR violence) AND (Surgeon OR Resident OR Intern OR physician OR doctor OR “general practitioner” OR “health care” OR Nurses OR Clinicians) AND (workplace OR hospital) AND (”risk factor” OR determinants OR predictors). This keyword string was searched on electronic databases such as PubMed and Google Scholar. Hand searches and contacting researchers were not carried out as a part of this study.
Searches were limited to studies that were peer reviewed and published in English language journals in 11-year period between October 2009 and September 2020; being quasi-experimental, case-control, pre-post (longitudinal), observational cross-sectional, cohort, or randomized controlled in design. In order to gather uniform quantitative data, case reports, case series, reviews, opinion pieces or commentaries or editorials were excluded. It should be noted that bullying as a form of violence was also excluded as the review focused on episodes of patient led violence against healthcare professionals.
Data Extraction: Detailed evaluation was done for those studies that met the criteria for inclusion and exclusion. The characteristics that were evaluated from the identified studies included the author, country, study design, sample size and characteristics, predictors and the findings. Data extraction was carried out by researchers (SC and TK) and discrepancies were resolved through consensus or third-party adjudication (PR).
Studies were screened according to quality of evidence: studies reporting perceived causes and studies reporting factors.
Methodological Quality: Critical appraisal of the methodological quality was done by Johanna Briggs Institute Critical Appraisal Tools for quasi-experimental studies. To assess the quality of the studies a checklist of nine items (cause and effect, control group, outcome measurement, follow-up, statistical analysis) was used with marking 'Yes', 'No' or 'Unclear'. The methodological quality of the studies was independently assessed by two reviewers (SC and TK). Any disagreements were resolved through consensus or third-party adjudication (PR). A risk of bias graph was formulated using Cochrane Collaboration Tool Review Manager Version 5.3 for randomized control trials included in this review. Due to heterogeneity in the study design and outcome measures, quantitative synthesis (meta-analysis) was not done as a part of the review.
Grading of Overall Quality of Evidence: The completed evidence of predictors was subjectively graded based on an adapted Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. It was graded on the following parameters: imprecision, risk of bias, indirectness, publication bias, and inconsistency. The rating of parameters was 'Yes', 'No' or 'Unclear'. The rating was done according to four quality categories as high, moderate, low and very low quality by two reviewers (SC and TK). Disagreements were resolved through consensus or third-party adjudication (PR). The definition of the four quality categories is given in [Table 2].
Table 1: Studies reporting predictors of workplace violence against health-care professional
Table 2: Adapted Grading of Recommendation for Assessment, Development and Evaluation (GRADE) table for narrative synthesis of studies for moderate and high-quality Results After the initial keyword search a total of 1250 articles were retrieved. All the titles and abstracts were assessed and the duplicates were removed, from those 145 full texts were assessed for inclusion in the review. A total of 46 studies were included (presented in [Table 1] and [Figure 1]).
Study Characteristics: Studies on predictors of WPV originated from 26 countries and a majority of them were from China (8), Iran (4), United States (4), Pakistan (3), and Italy (3). Of these studies, ten studies each were reported from teaching hospitals,[12],[14],[16],[17],[18],[20],[24],[27],[32],[33] public hospitals[15],[16],[20],[28],[29],[32],[33],[37],[47],[49] and healthcare facilities,[9],[22],[23],[28],[34],[35],[37],[40],[41],[48] seven from emergency departments[12],[13],[21],[30],[31],[38],[43] five from psychiatric units[8],[19],[35],[42],[46] four from multi-centric healthcare units,[10],[33],[37],[39] two from nursing,[14],[45] and one from ambulance service.[30] Majority of the studies were cross-sectional surveys.[10],[11],[13],[15],[16],[17],[18],[19],[20],[21],[22],[23],[24],[25],[26],[27],[28],[29],[30],[31],[32],[33],[34],[35],[36],[37],[38],[39],[40],[41],[42],[43],[45],[46],[47],[48],[49] Three studies had longitudinal design and other two studies had retrospective case-control design[8] and mixed-method design.[9] A consensus was observed among the studies in defining 'WPV' as the incidents where staff was abused, threatened, or assaulted in work-related circumstances involving a direct or indirect challenge to their safety, well-being, or health.
Healthcare professional characteristics: There was significant variation in sample size from 112 to 15,970, with female preponderance (range, 100% to 32.1%)[10],[11],[50] and the mean (standard deviation [SD]) of age (range, 41.5 [11.7] to 55 [10.8]). All the participants had basic diploma or undergraduate degree with mean experience (range, 2.95 [0.77] to 18.47 [11.93])[12],[13] except one study which recruited first-year nursing students.[14] Participants included were healthcare professionals in 11 studies,[15],[16],[18],[22],[24],[28],[35],[38],[40],[41],[44] nursing professional in 23 studies[9],[10],[11],[12],[13],[14],[17],[18],[19],[20],[23],[24],[26],[27],[30],[31],[39],[41],[42],[45],[46],[47],[49] followed by doctors in 11 studies;[12],[21],[25],[26],[29],[32],[33],[36],[37],[43],[47] doctors and nurses in 4 studies[12],[26],[40],[47] and 1 in ambulance personnel.
Methods of data collection and statistical analysis
Most of the studies collected data using self-administered questionnaires to assess the different domains of WPV: sociodemographic characteristics, source of violence, frequency and type/form of violence (physical or verbal abuse), perceived reasons, reporting of incidents, and previous experience with violence. A few studies administered validated scales/tools such as Violent Incident Form,[15] Violence in Healthcare setting,[16] Copenhagen Psychosocial Questionnaire,[17] and Perception of Aggression Scale.[18] The significant predictors were analyzed using descriptive statistics such as proportions in most of the studies and some studies used inferential statistics such as logistic regression analysis, Chi-square test, ANOVA, and Student's t-test.
Methodological quality of studies: After critical appraisal of cross-sectional studies [Table 3], cohort studies and case-control study, the risk of bias graphs were obtained (depicted in [Figure 2]). Cross-sectional studies (n = 44) had low risk of bias for most of the items, except two items identification of confounding variables and strategies to address them. Cohort studies (n = 2) were rated as low risk of bias [Figure 3] for the following items: recruitment of two groups from similar population, similar measurement of exposure, assessment of outcomes and complete follow-up from valid and reliable measures [Table 5], except for item strategies for incomplete follow-up, which was rated as unclear risk of bias.[19] Case-control study had fair methodological quality with unclear risk of bias for items such as reliable and valid measurement of exposure and outcome and duration of exposure[8][Table 4].
Findings Indicative Evidence: The predictors of WPV reported in these studies can be categorized as: patients and professional-related factors, work-related factors (job type, management, and support) and societal factors (summarized in [Table 6]).
Table 6: Adapted Grading of Recommendation for Assessment, Development and Evaluation (GRADE) table for narrative synthesis of studiesPatient-related factors: Studies have identified patients[12],[14],[50] and their attendants[16],[18],[20],[50] especially partners and siblings[18] as the main source of WPV. Some patient-related demographic factors such as being male,[13],[19],[21] lower education,[16],[22],[23],[24],[25],[26] and high social status[25] were significant predictors of WPV. Age of the aggressor and residence had contradictory evidence. Patients seeking medical help for violent behavior and psychiatric illnesses,[14],[15],[27],[28],[29],[30],[31],[32],[50] chronic disease,[25] and pain[14] were more commonly violent against healthcare workers. Patients having low impulse control[14],[19],[30] or under the influence of drugs and alcohol[14],[16],[17],[19],[21],[22],[24],[27],[29],[31] also were more likely to initiate WPV. Some acute situations such as patients' dissatisfaction with the service[13],[14],[17],[19],[27],[28],[31],[33],[50] or denial of service,[33] patient's death,[16],[24],[25],[33] legal cases,[16],[24] and no visitors[16] or too many visitors[25] were commonly reported causes of WPV. The overall effect for patient-related factors was found to be [OR: 1.64 (CI: 0.76-3.55)], the most significant predictor was patients with mental health issues. Substance abuse and patient dissatisfaction were not found to be significant (Supplementary File).
Professional factors: Professional-related factors such as demographics, nature of work, and personal characteristics were associated with WPV. Some of the demographic variables such as being female,[11],[14],[15],[34],[35],[36] being single child, lower education level (diploma or bachelor's degree),[17],[26],[28],[31] lower income status of physicians,[37] and higher for nurses[33] were positive predictors of WPV. Other factors such as race, marital status and age were inconsistently reported. Among healthcare professionals, nurses were more prone to episodes of WPV.[9],[11],[15],[22],[23],[28],[37],[38],[39],[40],[41] Nurses are more frequently in contact[14],[15],[42] with patients to re-enforce compliance to the treatment[19] which puts them at a high risk for WPV. Also, shift workers report more incidence of WPV during the night shift.[21],[22],[26],[37] Some personal characteristics of healthcare workers such as indifferent/rude/anxious behavior[10],[11],[13],[19] inability to de-escalate patient's feelings,[19] poor co-worker relationship,[10],[17],[27],[28] and worrisome behavior were considered as common causes of WPV. Previous experience of WPV was a positive predictor of future episodes.[16],[26] The overall effect of profession-related factors was found to be [OR: 1.47, (CI: 0.86-2.49)], three studies out of seven reported that being a nurse was a significant predictor of violence in healthcare settings. Only one out of five reported that shift workers were at increased risk. Being a female worker was also found to be a significant profession-related factor (Supplementary File).
Organizational factors: Studies suggest environmental factors such as operational setting, administration and management practices, and organizational support are associated with WPV. WPV episodes were more prevalent in hospitals.[21],[28],[26],[35],[34] Certain departments such as emergency[11],[15],[18],[23],[36],[38],[39],[41],[44] were at higher odds of experiencing WPV, followed by inpatient geriatric care.[15],[34],[37] Waiting area for patients was a prime location for WPV.[30] The denial of a patient's request for admission was reported as a cause of WPV. The period between admission of the patient and establishing diagnosis also witnessed episodes of WPV.[29] Some poor administrative practices during the management of patients which are commonly reported as causes of WPV include the following: long waiting hours[13],[22],[24],[25],[27],[29],[30],[31],[50] and uncomfortable waiting rooms,[27] overcrowding,[13],[27] clustering of high-risk patients,[16],[21],[24] limited resources such as medicines,[13],[33] equipment[25] or staff, high workload, miscommunication,[13],[14],[21],[27],[33] violation of visitor hours,[29] and prolonged stay after discharge.[16] Factors such as assisting patients in waiting rooms, communicating with patients, and professionals' ability to compromise are considered as preventive factors for WPV.[30] The overall effect of organization-related factor was found to be [OR: 0.90, (CI: 0.22-3.66), only one study out of five reported that emergency setting is most prone to the violent attacks. Other factors such as increased workload, waiting are setting, delay in treatment, and Government hospitals were reported to be significant organization-related factors (Supplementary File).
Organizational support to create a safe work environment was seen as a preventive factor for WPV. Some of the measures reported to prevent WPV against healthcare workers were: prioritization of WPV, motivation to report WPV, compliance to safety protocol, penalty for aggressors, support from superiors, and employee participation in training against WPV.[15],[21],[24],[32],[45]
Societal Factors: Some societal factors which contribute to episodes of WPV are negative media image of the healthcare industry, patient's distrust in 'expensive and inaccessible healthcare system,' and lack of policies to protect health workers against WPV.[13],[14],[19]
Most promising predictors: Confirmatory evidence: From the graded overall evidence, most predictors were categorized as very low- and low-quality evidence. A few predictors such as the patient's history of mental health issues, being male, experience less than five years, long employment duration, psychiatry ward, and professionals working in shifts especially evening shifts were graded as moderate quality evidence. Being a nursing officer was rated as high-quality evidence, hence contemplated as the most promising predictor in this review.
Publication bias: Funnel plots have been formulated for three factors, viz, shift worker, nurses from professionals, and emergency department from organizations (Supplementary File). As per the forest plot, these were the most promising predictors. Egger's test was used to compute the publication bias of the above-mentioned predictors, none of them were found to have significant associations (p > 0.05) which indicates that none of these have any publication bias.
DiscussionPredictors of WPV were assessed in this systematic review in different healthcare professionals working in hospital-based settings across the globe to identify factors which can be integrated for effective prevention and mitigation strategies. The review has a majority of cross-sectional studies[10],[11],[13],[15],[16],[17],[18],[19],[20],[21],[22],[23],[24],[25],[26],[27],[28],[29],[30],[31],[32],[33],[34],[35],[36],[37],[38],[39],[40],[41],[42],[43],[45],[46],[47],[48],[49] providing indicative and confirmatory evidence. Most of the studies[9],[10],[11],[12],[13],[14],[15],[16],[17],[19],[20],[21],[22],[23],[24],[25],[26],[27],[28],[29],[30],[31],[32],[33],[34],[35],[36],[37],[38],[39],[40],[41],[42],[43],[44],[45],[46],[47],[48],[49] used a questionnaire-based survey to collect data from middle aged nurses and other healthcare professionals working in hospital facilities, especially in the emergency department. The reported predictors of WPV were categorized as patient, professional, environmental, and societal factors.
Patient-related predictors help to identify high-risk patients capable of committing violence against healthcare professionals. In our review, the history of mental health issues in a patient was identified as a high-risk predictor of WPV. The history of mental health issues was defined as patients suffering from depression, anxiety, stress and psychosis, who were capable of self-harm. Since these patients are mostly found in psychiatric care, we also found psychiatric care units as another predictor associated with episodes of violence. Similarly, another systematic review emphasized that almost one in five patients admitted to acute psychiatric units may commit an act of violence toward healthcare professionals. The review also highlighted additional characteristics such as gender as male, history of schizophrenia, substance abuse, and violence would increase the probability of assaults by manifolds.[51] Patients with psychiatric illness are affected by episodes of confusion or altered mental states, hallucinations, or delusions, which limits their abilities to self-regulate emotions and increase impulsive outlet of aggressive emotions as episodes of assaults on healthcare personnel.[52] Mental health nurses are at three times greater risk of assaults than other healthcare personnel.[53] Some of the personality-related factors of the care staff such as nurses and ward staff with 'bad attitude”, rudeness and sarcasm were related to higher episodes of violence, whereas care staff with “calm attitude”, empathy, and good communication skills was associated with significantly lower chances of violent episode in a psychiatric in-patient setting.[54] Other professional-related demographic variables such as age, gender, education, and experience have been inconsistently related to odds of assaults and abuse in our review.
We also found out that nursing professionals were identified as the most vulnerable group of professionals facing WPV. These findings might be due to higher representation of the nurses in the studies included in this review, and the higher nurse to patient ratio as compared to doctor to patient ratio. Similarly, a meta-analysis also identified that nurses followed by physicians were more commonly abused than other healthcare professions.[1] We also found that both genders were at equal odds of experiencing WPV. Although, the nature of violence experienced was different amongst the two genders. Our findings were concurrent with a meta-analysis which found that female nurses were more verbally abused and male nurses were more physically abused.[49] Besides, all professionals who were young and working longer hours (>40 hours per week) in shifts were at a higher risk for any type of WPV.[1]
In our review, we found that professionals who had less than five years of experience and worked in shifts, especially evening shifts, were more prone to episodes of violence. Young professionals at the earlier stages of their career have to learn a diverse set of skills from different departments, which would make them shift their base departments frequently. They might be expected to work the odd shifts to accommodate their senior professionals in the departments. Other factors that could contribute to greater odds of physical violence during evening or night shifts are limited staff, staff exhaustion, frustration, increased patient contact, longer working hours, and less surveillance.[54]
On the contrary, we also found longer employment duration also led to greater prevalence of violent episodes. Although, the working definition of longer employment varied across the studies. For the purpose of better understanding, we defined greater than 10 years of experience as longer employment duration. One plausible explanation for our findings could be that senior professionals have greater responsibilities, heavier workload, and involvement in patient care. Senior doctors may have increased patient contact in their daily practice, which may impact quality of care.[37] Patients also have greater expectations from the senior professional, which if left unmet, might provoke the assault.[47]
Implications: To our knowledge, this is the first review which has tried to analyze the predictors of WPV for all the healthcare professionals working in a hospital-based setting. A number of perceived causes of WPV were identified, which can be used as a checklist by the professionals to identify predictors specific to their professional setting and nature of work. Some of the predictors which are applicable to all the healthcare professionals were also identified which can be used as the fundamental basis to develop and implement prevention and mitigation strategies.
Limitations: The interpretation of the results requires careful consideration of certain aspects, Firstly, the number of case-control and cohort studies were limited in the literature review. Apart from this, only studies with English language were selected which could possibly lead to exclusion of studies from different languages. Secondly, we looked at only two databases for identifying the studies which could lead to possible exclusion of available literature on other databases. Lastly, the study has been done as per PROSPERO guidelines but we could not register this protocol before the initiation of the study.
ConclusionsThis systematic review has identified being a nurse as the most important predictors of experiencing WPV followed by other factors such as a patient's mental health, professional-related factors (gender, experience), and environmental factors (psychiatric unit and evening shifts). This review has also reiterated the fact that an episode of violence is an interplay of multiple factors. Yet, an understanding on the utilizing different predictors to develop an implementable strategic framework for prevention and mitigation of these episodes is missing, which can be taken up by future studies.
Key messages
To our knowledge, this is the first review which has tried to analyze the predictors of workplace violence against healthcare professionals working in a hospital-based setting.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
Supplementary File
References
Comments (0)