The Demand-Control model and work-related threats and violence: Short- and long-term associations

1.1 Risk factors for work-related violence

Work-related violence is a complex phenomenon, with no single cause being sufficient to explain why some employees are exposed to work-related violence and others are not [8]. Studies have found that risk factors at the individual level such as being young, having less job experience and holding a lower rank in the organization are associated with higher risk of work-related violence [9, 10]. However, these individual factors seem to occur as a result of the interaction between the persons involved and factors in the social context [8]. Accordingly, several studies have explored risk factors at the workplace level, focusing on specific work environment factors that could increase the risk of work-related violence and threats. Specifically, workplace factors including high demands, high work pressure, low role clarity, poor social climate and social support, high emotional demands, low levels of involvement at the workplace, lower leadership quality, and role conflict have been found in association with elevated risk of work-related violence and threats [11–14].

An important limitation of these studies is that they are mainly cross-sectional, with only a few adopting a longitudinal design. For instance, a follow-up study including 3011 participants found that high quantitative and emotional demands, as well as high role conflict, were associated with heightened levels of work-related threats and violence [11].

Another important limitation of previous studies is that they have addressed an extensive list of work environment factors potentially associated with work-related violence and threats [11, 14]. Statistically, this is problematic because including many factors in a single regression model may lead to multicollinearity, which entails unstable and uninterpretable estimates. Theoretically, this approach is problematic due to the lack of a solid theoretical model on which to test associations and to further theoretical development. Using theory-driven models of the psychosocial work environment such as the well-established Job-Demands-Control Model [15], may therefore enhance our understanding of the specific working conditions that can increase the risk of work-related violence and threats.

1.2 The Job Demand-Control model

The Job Demand-Control model (JDC) distinguishes between two crucial aspects of the job, namely psychological demands (work load, unexpected tasks and job-related interpersonal conflicts) and job control (ability to control own work situation, autonomy and skill discretion). The model posits that the combination of high job demands and low job control (the strain hypothesis) results in the most adverse reactions in terms of psychological strain and physical illness [16]. However, job control gives employees the opportunity to have decisional authority on how they handle the required work, as well as to restructure the pacing, the timing and the methods used in accomplishing the work tasks [17]. Thus, high job control is expected to moderate the negative effects of job demands on employee health and well-being.

Many studies have applied the JDC model to a broad array of outcomes such as mortality, angina pectoris, mental strain and psycho-somatic symptoms. As pointed out by Baillien et al. [18], behavioural outcomes of the model are rare and have only been addressed in terms of work-performance and health-related behaviour. Some studies responded to this gap by testing the strain hypothesis of the JDC model in relation to workplace bullying, finding that high strain was associated with targets’ reports of bullying and that job control moderated such association [18, 19].

1.3 The JDC model and work-related violence and threats

While the previous mentioned studies highlight the predictive ability of the JDC model in relation to workplace incidents such as bullying, it is still unknown whether the model is capable to predict other forms of aggressions, including work-related violence and threats.

Both empirical findings and theoretical arguments support the integration of the JDC model into the research on the antecedents of work-related violence and threats. Empirical studies have shown positive associations between high workload and work-related violence. For instance, victims of work-related violence have been found to often experience high qualitative and emotional job demands and low control over own work situation [11, 14].These work characteristics are at the core of the JDC model. From the theoretical viewpoint, a link between the JDC model and risk of work-related violence can be expected because situations of high job demands and low job control may increase work-related strain (for a review, see Häusser et al. 2010 [20]) which, in turn, may decrease job performance in terms of disrupted continuity of care, increased errors and undermined quality of care and service [21, 22]. Care recipients and service users (e.g., patients, client, pupils, and prisoners) may, in turn, react in violent and threatening manners in the face of errors, mistakes or blunders that occur in conditions of heavy work load, high work pressure and low job control [18, 23].

To date, only a few studies have used the JDC model to explore the relationship between the work environment and work-related violence and threats. Two studies found that high demands were a significant predictor of threats and emotional abuse; however, job control was not associated with work-related violence or threats, and no interactive effects were observed [13, 24]. The validity of these findings is, however, limited due to methodological shortcoming such as small samples sizes, low response rates, the use of a cross-sectional design and the inclusion of nurses and midwives exclusively, which makes generalization difficult.

1.4 Study aim and hypotheses

The aim of the present study is to examine whether high levels of work-related work demands and low levels of job control (i.e., job strain) affect the report of work-related violence and threats in four high-risk occupational sectors. Specifically, the following hypotheses were tested:

High quantitative demands at work are positively related to reports of work-related violence and threats.

2.1 Participants

The study sample is based on a Danish cohort established in 2010 [4], with follow-up surveys carried out in 2011 and 2015. The associations between work demands and job control measured in 2010 and work-related threats and violence measured in 2011 are reported elsewhere [11]. In this study, we want to examine the long-term associations between work demands and job control measured in 2011 and work-related threats and violence measured in 2015. We included those employees who participated in the 2011 and 2015 waves of the study.

The cohort consisted of employees working in psychiatric wards, in the elder care, at special schools (schools for pupils with severe developmental and intellectual disabilities) and in the Prison and Probation Service. The participants employed in psychiatric wards, elder care homes and at special schools received and filled out paper-and-pencil questionnaires during a planned meeting at the worksite. A web-based questionnaire was used for participants in the Prison and Probation Services –a decision made by top management. The invited participants were persons employed in jobs involving contacts with clients and with no record of absence spells of more than three weeks prior to the time the questionnaires were distributed. The last criterion was to ensure that employees were sufficiently aware of the present psycho-social work environment at their workplace. The procedure for data collection is described in detail elsewhere [4]. In the questionnaire, the participants were informed that participation in the study was voluntary and that the participants could withdraw from the study at any time. According to Danish Law, approval by the Ethics Committee and written and informed consent is not required in questionnaire-based and register-based projects.

In 2010, 5,333 employees participated. Altogether, 3,751 participated in 2011. In 2015, the data collection was based on invitations sent to all participants from 2010 who were still living in Denmark. Using their civil registration number, we were able to contact them by mail, regardless of whether they were still employed at their original workplace. A total of 3,486 employees participated in 2015.

In this study we included all those participants who filled out the questionnaires both in 2011(N = 3751) and in 2015 (N = 2,675) given a response rate of 76.7%.

2.2 Questionnaires

2.3 Statistical analysis

The data did not meet the statistical assumptions of normality and homoscedasticity required for multiple linear regression [28]. We therefore applied hierarchical logistic regressions and calculated odds ratios, along with their 95% confidence intervals (CI), for the associations of quantitative demands and job control with work-related violence and work-related threats. The two scales were dichotomised by contrasting those exposed to work-related threats and violence versus employees not-exposed.

Quantitative demands and job control were continuous scales.

In the first block of the regression analysis, we entered the control variables (sex, age, working time spent contact with clients, and, only for the longitudinal analysis, baseline exposure to work-related threats and violence). In the second block, we entered the quantitative demands and job control, while in the final block we added the interaction between quantitative demands and job control.

4.1 Strengths and limitations

The present study has some important strength. First, the response rates were high and above the mean for organizational surveys [37]. Second, we examined participants from four sectors, which may increase generalizability. Finally, the study is based on a longitudinal design.

Yet, the results of the present study should be considered in the light of some limitations. First, in three sectors (eldercare, special schools and psychiatric wards), the participants were not representative of the respective population, since the data derived from convenience samples of employees coming primarily from two regions in Denmark. Even though the total sample size was large, these organizations were recruited using a non-random sampling method. Second, as some workplaces refused to participate, we cannot rule out selection bias, which may reduce the external validity of the findings. Third, modes of data collection were different across the four sectors, as web-based questionnaires were used in Prison and Probation Service and paper and pencil questionnaires in the other sectors. This may have introduced selection bias. If only employees showing interest in the topic responded, this may have inflated the associations, especially in the Prison and Probation Services wherein the lowest response rate was observed. Fourth, only healthy employees participated in the current study; those on sick leaves, including leaves due to work-related violence and threats, did not participate, which may lead to an underestimation of the associations between psychosocial work environment factors and work-related threats and violence. Fifth, the data came entirely from self-reports, which may introduce mono-method bias due to unmeasured third variables [38]. Final, the results must be interpreted with caution as there may be specific sector associations which may be blurred when the four sectors are analysed together. For instance, the possibility for high control may be different in schools and prisons.