Introduction
“Third-party violence” in the workplace is a term used to refer to threats and physical and psychological violence (e.g., verbal violence) by third parties such as customers, clients, or patients receiving goods or services1). Possible consequences for individual workers, teams, and organizations include impaired psychological well-being, mental and physical illness, absenteeism, impaired performance, and higher turnover2, 3). Initiatives to address workplace violence come in at the political level (regulations, agreements, guidelines), organizational level (corporate policies, workplace risk assessment, and control), and the level of the individual worker (training, case management, and support).
The health-care sector is the sector with the highest rate of third-party violence against workers. In Europe, workers in the health sector are eight times more likely to have experienced a threat of physical violence than workers in the manufacturing sector3). In the US, the health-care sector accounts for 53% of nonfatal injuries due to workplace violence4) and health-care workers are the first occupational group to report violent victimization at work5). Emergency and psychiatric departments are the most affected.
Many previously published studies have taken an epidemiological perspective, seeking to determine and quantify the risk factors, modalities, and consequences of the problem6–8). This objective approach fits in with the legal framework obliging employers to ensure the health and safety of their employees. Occupational risk of violence has to be assessed9): the efficacy of measures undertaken has to be measured with adapted indices10). This approach, while necessary, neglects the individual subjectivity inevitably involved in assessing and adapting to any situation of violence. Qualitative research protocols have therefore been developed to explore health-care workers' actual experience, usually by adopting the individual point of view of nursing staff8, 10). Topics thus explored include the causes, consequences, and meaning attributed to violent events. These studies have stressed the importance of the interpretation of each such event as construed on a case-by-case basis by the health-care workers involved11–15), helping account for the phenomenon of underreporting of violent incidents. Few studies have adopted a broader perspective seeking to understand violence in the hospital in its social and organizational context16–18).
In France, acts of violence in hospital have been recorded since 2005 by the National Observatory of Violence in Healthcare Settings (Observatoire National des Violences en Milieu de Santé)19), using a 4-level typology of seriousness based on the French Penal Code. Data for 2012 are shown in Table 1. Emergency, psychiatry, and geriatric departments are the most badly affected.
Table 1.
Typology and proportion of acts of violence in France in 2012
a
| Level of seriousness |
Types of victimization |
Report rate (%) |
| 1 |
Insult and provocations without threat (offensive speech of a discriminatory or sexual nature) |
28% |
| 2 |
Physical threat |
20% |
| 3 |
Physical violence (attack, hustling, spitting, hitting), armed threat, sexual aggression |
51% |
| 4 |
Armed violence, rape, criminal act (murder/manslaughter, intentional violence causing mutilation or permanent disability, kidnapping, illicit detention, etc.) |
1% |
a Total of 11,344 incidents reported by 12% of the institutions involved in the Observatory.
Increasing use of emergency departments leads to increased user waiting times and time pressures on health-care workers. This phenomenon was observed in a specialized ophthalmology emergency department of a university hospital in the Rhône-Alpes region of France, where the number of admissions increased 5-fold in 20 years, from 4,000 in 1984 to 23,000 in 2012. In parallel, the rate of acts of violence and antisocial behavior reported by personnel increased, with both medical and nursing staff demanding that this issue be addressed. In this context, an interventional epidemiological study, PREVURGO, assessed 5 interventions with the number of acts of violence and antisocial behavior reported by personnel as the efficacy criterion20).
Given the influence of representations on the reporting and underreporting of acts of violence, a qualitative pilot study was conducted in personnel with the objective of exploring and describing the ophthalmology emergency department personnel's experience of acts of violence and antisocial behavior.
Methods
A grounded theory qualitative research approach21) was adopted in order to capture interactions between the various agents in the real-life working context and to take account of the points of view expressed by the health-care workers.
Context
The study was carried out in the ophthalmology emergency department of a large university hospital with an urban catchment of 1.3 million. The number of emergency admissions per year was close on 23,000, for a daily average of about 60 and a peak of up to 100. During daytime hours, reception is conducted by an administrative agent and a nursing auxiliary who complete the necessary paperwork and dispatch the patients to two-person teams comprising a nurse and an intern. Between 5 p.m. and 8 a.m., patients have to go to another building to complete admission paperwork. A total of 26 people worked in the emergency department: 11 nurses (including 1 head nurse), 8 interns, 5 nursing auxiliaries, and 2 administrative agents (receptionists). Other nurses and receptionists worked there on an occasional basis (weekends and public holidays). A senior physician was on call and was brought in when surgery was required. The premises comprised an administrative reception office (during daytime hours), 2 separate waiting rooms, a nursing office, a consultation cubicle for interns and a dressings room.
Study subject sampling
The sampling design sought to include health-care workers in each occupational category: administrative, paramedical, and medical. Given the small numbers involved, all team members were included so as to have as complete a description as possible. Two persons (one nurse and one intern) declined to participate, pleading lack of time. The total number of participants was thus 30: 15 nurses (including the head nurse), 5 nursing auxiliaries, 7 interns, 2 receptionists, and 1 senior physician (see Table 2).
Table 2.
Participant characteristics
| Code |
Gender |
Age |
Occupation |
| NA.01 |
Female |
35 |
Nursing auxiliary |
| NA.02 |
Female |
48 |
Nursing auxiliary |
| NA.03 |
Female |
56 |
Nursing auxiliary |
| NA.04 |
Female |
50 |
Nursing auxiliary |
| NA.05 |
Female |
47 |
Nursing auxiliary |
| NUR.01 |
Female |
60 |
Nurse |
| NUR.02 |
Female |
40 |
Nurse |
| NUR.03 |
Female |
40 |
Nurse |
| NUR.04 |
Female |
58 |
Nurse |
| NUR.05 |
Female |
55 |
Nurse |
| NUR.06 |
Female |
57 |
Nurse |
| NUR.07 |
Female |
53 |
Nurse |
| NUR.08 |
Female |
46 |
Nurse |
| NUR.09 |
Female |
50 |
Nurse |
| NUR.10 |
Female |
53 |
Nurse |
| NUR.11 |
Female |
48 |
Nurse |
| NUR.12 |
Female |
35 |
Nurse |
| NUR.13 |
Female |
52 |
Nurse |
| NUR.14 |
Female |
42 |
Nurse |
| NUR.15 |
Female |
54 |
Head nurse |
| INT.01 |
Male |
25 |
Intern |
| INT.02 |
Female |
26 |
Intern |
| INT.03 |
Female |
26 |
Intern |
| INT.04 |
Male |
26 |
Intern |
| INT.05 |
Female |
27 |
Intern |
| INT.06 |
Male |
31 |
Intern |
| INT.07 |
Female |
30 |
Intern |
| MED.1 |
Female |
36 |
Senior physician |
| REC.1 |
Female |
53 |
Receptionist |
| REC.2 |
Male |
61 |
Receptionist |
NA, nurse auxiliary; NUR, nurse; INT, intern; MED, senior medical doctor; REC, receptionist.
Three investigators dealt with data collection: CDA (medical intern, new to research work), JBF (occupational physician, experienced in research), and PS (social and occupational psychologist, experienced in research). They were equally involved in the entire data collection process and met regularly during the study period to exchange ideas. Data collection was performed with different modalities to ensure data source triangulation22): field observation in the department, participation in meetings, individual interviews, violence report forms, and patient satisfaction questionnaires.
1) Nonparticipant observation
In a first step, the investigators performed nonparticipant observation, in June–July 2012. Both users (at reception and in the waiting room) and staff (nurses' office and interns' consultation cubicle) were observed during sixteen observation periods of 3 to 4 hours each, for a total 52 hours over 6 weeks, focused on day- and nighttime department functioning. Staff, but not users, were informed of the observation. The investigators thus gained an overview of the field and were able to identify several critical situations in the actual setting.
2) Semistructured individual interviews/initial conceptual framework
A semistructured individual interview guide was drawn up based on the field observations and the literature on violence in emergency departments. The International Labor Organization's interactive model of violence at work23) was adapted to construct a conceptual framework to organize the data. This particular model was selected because it includes both the individual and organizational factors identified in the observation phase, and it was modified to include the stages of adaptation to violence. The conceptual framework, presented in Fig. 1, was an interactive (relations between various persons), contextualized (work organization), and process-based model (phases of patient management, adaptation strategies, consequences of violence).
The interview guide was based on this conceptual framework and tested with the other investigators. Some questions were reformulated to improve clarity in the light of the first interviews. Table 3 presents the main topics. The guide was adapted for interviewing the senior physician and the two administrative agents. Interviews were conducted equally by the three investigators (CDA, JBF, PS) during October 2012.
Table 3.
Main topics of the interview guide
| Sociodemographic data (age, experience in the unit, previous experiences) |
| Job description |
| Examples of incidents (last incident, most usual incident, most frightful incident) |
| General factors liable to contribute to aggressive behavior (waiting time, waiting room, degree of emergency, alcohol consumption, day or night shift, etc.) |
| Individual strategies in case of violence (immediate and secondary) |
| Joint strategies in case of violence |
| Health impact |
| Occupational impact |
| Solutions to violence |
3) Field notes and investigators' logbook
The three investigators took field notes during the observation phase. These were fed into the process of drawing up of the interview guide and helped contextualize the subsequent data analysis. CDA and JBF kept a logbook during the interview phase with analyses to guide the interpretation of results and also to keep a trace of each investigator's individual reflections.
4) Violence report forms
All staff could report violent incidents to the hospital's security service, using a standardized report form. To corroborate findings by triangulation, the emergency department's report forms for the previous three years were collected and analyzed in the light of the field observation and interview analysis.
5) Patient satisfaction questionnaire
All patients could fill out a satisfaction questionnaire at discharge. Those filled out by ophthalmology emergency patients during the 16 months preceding the start of the study were collected; responses to open questions were typed out and analyzed. In all, 364 questionnaires were analyzed (78 for 2013 and 276 for 2012).
Data analysis
1) Interview analysis
Interviews were recorded, transcribed, and analyzed qualitatively by topic. Two investigators (CDA and JBF) undertook this analysis. As a preliminary step, transcripts were rendered anonymous (persons' names) and checked for accuracy. Analysis then proceeded deductively according to the interview guide topics. In the first step, both investigators precoded three interviews and then checked the other investigator's coding; discrepancies were discussed and resolved by consensus. The precoded interviews were then entered into MAXQDA software v11 (VERBI GmbH). The second step consisted in clarifying or altering the initial analytic categories, creating new categories founded on the field observation data. The final step was to reduce, summarize, and correlate data by identifying meta-categories. The results were discussed by all the investigators and presented to two key informants for validation.
2) Analysis of patient satisfaction questionnaires
Satisfaction questionnaires were first analyzed in Excel and then in the MAXQDA v11 software. Five topics were distinguished.
3) Triangulation
The final results and their interpretation were discussed with the other investigators and modified accordingly. Interview analysis results were compared by triangulation with the field notes from the initial observation phase, the violence report forms, and the satisfaction questionnaires. From this analysis, a medium-range theory of staff experience of violence was formulated, which was grounded in the everyday occupational context.
Research ethics
The study was approved by the Lyon University Hospital Ethics Committee. Staff were provided information about the study collectively by three investigators (CDA, JBF, PS). An information sheet was delivered to each interviewee, who signed a consent form prior to being interviewed. All names and place names were replaced with codes to ensure anonymity.
Results
Coding led to 7 main topics broadly corresponding to the interview guide categories: acts of violence and antisocial behavior, contextual factors liable to induce violence, immediate individual reactions and strategies, collective strategies, violence reporting, health and occupational impact, and solutions.
Acts of violence and antisocial behavior
1) A wide range of different situations
A typology of acts of violence and antisocial behavior was drawn up with 6 categories of increasing seriousness: impoliteness, vulgarity, nonrecognition, insults, verbal threats, and aggressive gestures. The typology, drawn up inductively from actual data, differed from the regulatory typology based on the French Penal Code (Table 1). Impoliteness may be on the part of the patients or on the part of those accompanying them. Shouting was mentioned as a common manifestation of user discontent; other forms of impoliteness include slamming a door, entering the medical office without knocking, interrupting a consultation, answering a cell-phone during a consultation, or queue-jumping.
“When people ask us something, they don't say ‘Hello’ or ‘please’. They talk to us a bit like we're some kind of door-mat.” (NA.03)
“(…) there were three people who came into the consultation cubicle (…). So the doctor said, “Look, I'm busy examining someone. You can't just come in like that… You haven't even knocked or anything.” (NUR.11)
Vulgarity was considered use of bad language, often with sexual connotations, but not actual insults. Nonrecognition is an intention to be hurtful by denigrating the health-care worker's position, usefulness, competence, or commitment. Respondents mentioned often being called “bureaucrats” in an insulting manner.
“Then, in the evening, there was this patient who told me, when I was at the end of my tether, ‘Anyhow, you don't do anything. You're just a bunch of bureaucrats. It's a well-known fact’.” (NUR.11)
Insults, mostly sexist, were mentioned by more than half of the respondents. Only four health-care workers said they had never been insulted. All the others were unanimous that insults were common, repeated, and “came with the job”.
“He said “motherfucker” (or words to that effect), and then, I just gagged, completely gagged!” (NUR.01)
Verbal threats were mentioned less often. One intern reported death threats from an especially agitated patient. Aggressive gestures were reported by most respondents: between users or aimed at staff or equipment.
“There was this one raised his hand, and I was scared all the same, being alone like in the office.” (NA.03)
“Then he was throwing stuff in the waiting room—threw it all on the floor.” (INT.04)
2) Nurses on the front line
Acts of violence and antisocial behavior were mainly aimed at nursing auxiliaries and nurses, who were in the front line dealing with users. Interns and physicians were less often exposed, seeming to be protected by their status.
“(…) people who're aggressive toward us, as soon as they're in front of the doctor, then they're real teddy bears. Straight away [sigh] they become real doormats.” (NUR.11)
3) Continual background noise
The image that formed from the participants' responses and the observation phase was that of a continual background noise of impoliteness, punctuated by sudden flare-ups of aggression. This aggression was exclusively verbal (seriousness levels 1 and 2), without physical violence.
4) Incessant demands and complaints from the public
All respondents mentioned that work was systematically interrupted by incessant complaints from patients and those accompanying them about waiting time and departmental functioning. The general opinion was that these interruptions were more exhausting and disturbing than the acts of actual violence.
“They want to know how many people there are ahead of them, and I can't tell them because we don't take people in order of arrival but instead take them depending on how serious it is. (…) That's what's so hellish! Always having to justify yourself. I can understand it, but when it gets too much, that's what wears me out.” (NUR.06)
Local context factors inducing violence
1) Premises and administrative procedures
Administrative complexity (procedures varying depending on the day and time) was identified as a source of confusion and discontent for users. Cramped, uncomfortable, out-of-sight waiting rooms were identified as aggravating user impatience.
“They do the paperwork before they look after you!!!” (Patient questionnaire)
“Waiting room with nothing for entertainment (TV, etc.), and not very comfortable metal chair. How are you supposed to wait for seven and a half hours with nothing to do?” (Patient questionnaire)
2) Lack of information and friendliness
Analysis of the 354 satisfaction questionnaires found 72 user complaints claiming to have been badly informed or confronted by unfriendly staff.
“It's not the waiting that's so hard but that you get no explanation.” (Patient questionnaire)
3) Work organization
Disconnection between medical, paramedical, and administrative tasks was identified as a factor exacerbating the burden of work experienced by health-care workers and the waiting time endured by users. Incessant telephone calls interrupt nurses' work, increase user waiting time, and frequently involve verbal abuse over the phone.
4) Working relations between health-care workers
Lack of teamwork between certain health-care workers (nurses and nursing auxiliaries) sometimes increases workload and user waiting time.
“It depends on the colleague. (…) Some lend a hand, then others make themselves scarce.” (NA.05)
5) Staff shortages
Understaffing in the emergency department increases waiting time, leading to disbelief and anger on the part of users. This was reported by most respondents and frequently mentioned in user satisfaction questionnaires.
“An obvious shortage of personnel leads to unacceptable waiting times.” (Patient questionnaire)
6) Waiting times
Overall, all of the local context factors combine to increase waiting times, which can be up to 10 hours. It was generally agreed by respondents that user intolerance of waiting times and lack of information were the main underlying causes of violence. This was clearly borne out by analysis of the satisfaction questionnaires: 259 complaints about excessive waiting time out of 354 questionnaires.
“An emergency department where you wait 6 hours to get treatment is quite unacceptable.” (Patient questionnaire)
“The waiting times are a scandal!” (Patient questionnaire)
7) Outside context
The general lack of ophthalmologists was mentioned as a factor contributing to the overload of the emergency department. The mean time to an appointment in the community is more than 6 months, and many patients therefore come to the emergency department because they have no other real option, leading to inappropriate use of the department for nonurgent treatment.
Individual strategies for dealing with the public
Several types of health-care worker strategy, identified from the responses of workers and from observation, are presented in Table 4.
Table 4.
Individual strategies toward the public
| Proactive strategies |
| Introducing oneself |
| Taking the initiative (in waiting room or corridor) |
| Providing explanations (waiting times, catering facilities, etc.) |
| Leaving the office door open |
| Avoidance strategies |
| Avoiding contact so far as possible (waiting room assimilated to a “tribunal”) |
| Not introducing oneself |
| Closing the office door |
| Not answering questions |
| Reactive strategies |
| Self-control |
| Explanations (education) |
| Establishing the framework (being firm, self-assertion) |
| Humor |
| Passivity, avoidance |
1) Proactive strategies or avoidance
An important contrast was observed between certain health-care workers who systematically adopted personal proactive strategies intended to create an atmosphere of confidence and calm and others who on the contrary adopted avoidance strategies to minimize contact with the public.
2) Reactive strategies
Five types of behavior in the face of violence were distinguished from participants' responses and from observation. Self-control and control of the situation involved the ability to master one's own feelings and the other person's feelings during difficult interaction.
“And the more worked up you get, they more agitated you are for the rest of the day. (…) So we have enough sense to take it on our own shoulders. If you get insulted, well, you do. Doesn't matter.” (INT.04)
Explaining matters to users involved an “educational” attitude of informing them about how the emergency department works and the types of treatment.
“I try with discussion first of all, of course. Explain things to them, or tell them again (…).” (NUR.02)
Establishing a framework involved being firm so as to get control over a user who was being antisocial or impolite. Humor was one type of behavior mentioned by most health-care workers as a way of defusing difficult situations.
“I like to kind of get them to laugh at themselves; that often works. With a bit of humor, a lot of things go down better.” (NUR.03)
Finally, passivity was also mentioned as a strategy in itself, that is, cautiously letting the situation play itself out without making things worse.
“I didn't even answer. I just left him to it, and he went away.” (NUR.05)
Calling in the security department was mentioned by several respondents as a strategy of uncertain value, given the long time to intervention and the attitude of the security agents, who are often thought to take the user's side.
“The problem is, when you call in the guards, they come, they take the patient out into the corridor, talk, and then (…) more or less say the patient was right.” (NUR.01)
3) Negative thoughts and feelings
Most respondents reported a feeling of powerlessness. This was aggravated by the need to justify oneself to the users regarding the lack of resources and problems of hospital organization. The feeling of doing poor quality work and giving a poor image of public-sector hospitals was mentioned by several respondents, who felt they were failing to keep up professional standards.
“You're a bit ashamed, but you don't have a choice. Sometimes I feel like I'm in a factory more than a hospital department.” (INT.02)
Fear was not mentioned as being a frequent issue. It was a real enough feeling, but occasional, in the case of serious verbal aggression or physical threat. Overall, respondents reported not living in fear of violence at work.
Many other negative thoughts and feelings were expressed. These included comments about “wounded professionalism” and infringed dignity, mentioned in connection with interaction with especially vulgar or insulting users. Weariness and resignation were mentioned as being the result of repeated antisocial behavior. Reactions of exasperation or condemnation were expressed, sometimes to the point of indignation, revolt, and anger.
“When someone insults you, you don't appreciate it. I think it's a question of dignity. (…) It's not right, and it's something I don't like.” (NA.03) “So I was like, ‘no, it's unacceptable!’ I said ‘no, I can't. It's not possible’ I… I was wound up, angry.” (NA.04)
4) Defense mechanisms
Several defense mechanisms meant to minimize the unpleasantness of violence were identified. The most common were forgetting and habituation. Others were treating the situation as being ordinary or unimportant, generalizing it, getting some distance from it or mocking it.
“You have to forget or you won't cope.” (NA.03)
“I think I don't want to remember, in fact.” (NUR.07)
5) Awareness of one's own limits
Several respondents recognized their own limits and that reaching them was simply “human.” Such situations were described as not being able to contain oneself, getting more angry than one meant to, losing patience, or behaving badly. Respondents said they might come to the end of their tether with a patient or a colleague, effectively escalating the aggression.
“Because, like, sometimes, well, you know, you're only human. So some days you're less helpful, less patient, more tired, and you get annoyed faster than others (…) you fly off the handle.”(NUR.03)
6) Empathy for users
Despite the emotional demands induced by user behavior, almost all respondents said they could understand such behavior caused by excessive waiting. Users' “panic” about being on time to pick children up from school or anxiety about health were points that made their attitude easy to understand. Respondents mentioned “putting yourself in the patient's place,” thereby manifesting empathy.
“Personally, I put myself in their place, because there's nothing strange if you've been waiting 5 or 6 hours to be seen and the person's not nice to you (…). I can see how it gets on their nerves having to wait so long (…).” (INT.06)
Collective strategies
Few collective strategies to cope with violence were mentioned or observed. In some cases, nurses mentioned calling in a colleague to deal with an aggressive user. The head nurse's support was regularly mentioned. There was no formal debriefing arrangement following difficult situations; informal mechanisms such as talking between colleagues were mentioned. Thus a feeling of failure on the part of physicians and administration to fully acknowledge problems was stressed by several respondents (nurses and nursing auxiliaries).
Reporting antisocial behaviors and acts of violence
All personnel had the possibility of making a report, by hand or by computer, following any antisocial or violent event. Despite the reported frequency of acts of violence, however, less than half of the respondents said they had ever filed a report. Reasons for reporting varied: “moving on”, a question of principle, a wish to “beef up the stats” in the hope of thereby improving the situation, a wish to be acknowledged as a victim, or preventively as protection against legal action. In several cases, it took repeated instigation by the head nurse for the respondent to decide to file a report. In all cases, however, there was the idea of a certain threshold of seriousness for a report to be made.
“It's really the guy who's gone too far that you're going to report. The usual stuff, I'd say, you don't report it. It's true.” (NUR.07)
Those who had never filed a report pleaded lack of feedback, time, the complexity of the procedure, or ignorance of the possibility of reporting. A feeling that reports were pointless was mentioned by several health-care workers, underlining the lack of any positive feedback and the responsibility of the administration for the local conditions inducing violence (inappropriate premises and procedures; staff shortage).
“I don't make reports. There's no point. We aren't recognized.” (NUR.01)
“(…) the response I got made me tend to say to myself: what's the use?” (NUR.07)
“(…) is there like a figure, some kind of quota of malevolence, before you're entitled to whatever it is?” (REC.2)
Health impact
Overall, respondents claimed not to have work-related health issues. Some minor disorders were imputed to work by some of them, such as sleep disorder, stomach ulcers, or overweight. None reported having consulted a doctor specifically for an occupational health problem. Five respondents mentioned occasionally taking medication due to problems at work: hypnotics in four cases and gastric antacids in the other case. This generally reassuring picture needs to be set against the one case of a health-care worker who was visibly suffering work-related exhaustion when interviewed. This person took early retirement for health reasons, highlighting the sometimes dramatic socio-occupational impact of situations of workplace violence.
Impact on work
Almost all the health-care workers interviewed considered that acts of violence and antisocial behavior had an impact on the quality of their work, motivation, or relationship with patients.
“Yes, it's true. I get the impression I'm much less friendly with people than I used to be. It's true.” (NUR.07)
Despite all difficulties, respondents testified to pleasure in work, which left most of them with a positive feeling on balance. Apart from patient contact, they mentioned three points in particular that gave them pleasure and satisfaction: variety, learning new skills, and recognition on the part of some patients.
“In spite of everything, people still thank us. (…) Once or twice, say, they come back and say ‘Goodbye and good luck’.” (NUR.02)
“And they're usually delighted and apologize. And that, for me… it's a kind of… then I know I've done my job.” (INT.05)
Solutions
The solutions identified by the personnel to deal with violence matched the factors identified in the local context: reducing waiting time by increasing staff levels, improving the premises (waiting room comfort, confidentiality), changing signs, simplifying administrative procedures, and improving user information. There were no proposals for better coordination of medical, paramedical, and administrative task or for reinforcement of teamwork and debriefing, except by the head nurse.
Summary of results
On the basis of the field observations, the testimony of health-care workers and the patient satisfaction questionnaires, it was possible to list health-care system factors (Table 5) that reduce or worsen violence and its consequences in the department.
Table 5.
Health-care system factors improving or worsening users' antisocial behaviors and violence
| Worsening factors |
Reducing factors |
| Macro-level (economic context, medical demography, legislation) |
| Macroeconomic demand for productivity, entailing staff cuts |
Regulatory occupational health and safety obligations, requiring risk assessment |
| Lower number of community physicians, entailing increased used of emergency departments and longer waiting times |
National reporting system: National Observatory of Violence in Health Care Settings |
| Meso-level (hospital, health-care department) |
| Length and complexity of administrative procedures |
New health worker training policy |
| Uncomfortable premises, lacking visibility; direct access to medical and nursing offices |
Management vigilance regarding all reports |
| Lack of information on department functioning |
Management support of affected workers |
| Unreasonable waiting times |
|
| Overwork |
|
| Poor coordination between medical, paramedical, and administrative tasks |
|
| Poor teamwork |
Good teamwork |
| No formal debriefing, leading to under-acknowledgment of violence |
Informal debriefings between colleagues |
| Lack of feedback concerning violence reports, leading to underreporting |
|
| Individual level (health-care worker) |
| Avoidance strategies (defensive attitudes) |
Proactive strategies |
| Aggressive behavior |
|
Discussion
Several of the present findings agree with the existing literature. The contribution of organizational factors to hospital violence was previously mentioned with respect to the comfort of premises24), waiting times17, 25), information provided to patients and those accompanying them24, 26), and administrative problems17). The critical importance of health-care workers' behavior has been previously highlighted. Behavior tending to reduce violence has been described in terms of empathic communication17, 24, 26), early proactive interaction18, 26), and verbal and body language expressing respect and confidence17). Behavior liable to induce violence was described in terms of avoidance of12), nonengagement with patients18), lack of empathy18), and systematic distrust17). Likewise, previous reports attributed underreporting of acts of violence to lengthy reporting procedures25), lack of feedback and of institutional support17), and factors diminishing the perceived seriousness of the acts of violence12, 13).
Other findings differ from those previously published. The absence of reported acts of physical violence was doubtless due to the nature of the department studied here (ophthalmology emergency), as previous studies were set in general emergency departments. Participants insisted that users' constant enquiries and moving to and fro were more burdensome than the actual acts of verbal violence, with constant pressure constituting a continual background noise. The identification of certain shortcomings in task distribution (coordination between medical, paramedical, and administrative tasks) and teamwork (sometimes insufficient mutual assistance) as factors liable to contribute to violence is an original finding. In particular, the burden of nurses' administrative tasks and repeated interruption of their work (telephone, enquiries) were sources of fatigue impairing their ability to adapt to the aggressive behavior of users frustrated by lengthy waiting times.
Studying the functioning of the department strikingly revealed how a system of health-care may contribute to hospital violence via various factors related to general context (political and financial environment), organizational context (hospital functioning and work organization), and health-care workers' behavior. The light this sheds on the issue is in contrast with the political and media image of hospital violence as a symptom of worsening violence in general in society, and as a problem that requires drastic measures to systematically combat any sign of it (“zero tolerance”). This political and media spin corresponds to the authorities' need to ensure their missions of law and order and public safety. The symbolic aspect of the hospital and of health-care workers as resources devoted to the welfare of the sick feeds into the representation of violence in hospitals as an unacceptable injustice toward conscientious and devoted professionals. The present findings, however, suggest a less black-and-white vision of the responsibilities incurred. In other words, it is inexact and unjust to claim that the violence comes exclusively from patients and their accompaniers: there are a certain number of factors concerning the organization and personnel of the health-care system that also play a role. Macroeconomic deficit reduction policies as applied to the hospital translate into targeted cuts in health-care personnel, increasing time to consultation, overuse of emergency departments, lengthened waiting times, and difficulties in maintenance of premises, which are liable to exacerbate users' frustration and aggression. Defective work organization and team cohesion also contribute. All in all, each individual bears his or her own share of responsibility in terms of behavior liable to induce or exacerbate situations of violence.
This vision of shared responsibility for hospital violence is certainly less politically correct than a representation that implicates the actual aggressor alone. It puts in question certain public policy decisions, modes of hospital organization and management, and individual health-care worker's behavior. It corresponds, however, to realities actually observed and referred to in several studies16, 18, 24, 26). It also corresponds to published hospital violence prevention guidelines, which stress the need to analyze the risk of violence as close to the ground as possible in order to be able to identify solutions matching the various causal factors identified in work organization27, 28).
Study strengths and limitations
Several measures were implemented to ensure the validity of the results of this qualitative study22, 29). The grounded theory methodology and conceptual framework were adapted to the study objective and context. All but two team members were included in the sample; it is unlikely that important information was missed due to the two refusals. Triangulation of data sources allowed several sorts of information to be corroborated. Observation allowed the users' point of view to be taken into consideration as well as interaction with staff. The investigators' personal reflections were cultivated by the field notes and logbook. Interview transcription and analysis with the MAXQDA software ensured process traceability. Interobserver coding agreement was checked. The analysis of the results was discussed by the research team. In a feedback session, the head nurse and several team members assured the researchers that they recognized themselves in the reported results.
Study limitations lay in the fact that data relating to the hospital management's point of view were not collected to complete the analysis. It was not possible to make comparisons with the point of view of medical, paramedical, and administrative executives concerning work organization. User viewpoints also were not harvested directly; resorting to satisfaction questionnaire responses certainly impaired the richness of the information.
The results confirmed the relevance of the type of intervention recommended in the PREVURGO study20), with causal factors analyzed by staff members, in conformity with published guidelines27). The findings raised staff awareness of the importance of reporting violent acts so as to be able to assess the efficacy of the measures undertaken. The feedback session allowed experience with individual violence-reduction strategies to be shared and also contributed to better recognition of the problem within the department.
Given the importance of teamwork and task distribution in preventing violence, it is recommended that future studies focus on these aspects to identify the mechanisms involved in hospital violence. Despite the methodological and ethical difficulties, it would also be desirable to document user viewpoints directly.
