Risk Assessment of Workplace Violence Against Nurses: How Data Collection Methods Influence Results—A Swedish and Italian Cross-Sectional Study

TL;DR

This study shows that how data is collected strongly affects reported rates of workplace violence against nurses, more than cultural or organizational differences.

Contribution

The study demonstrates that data collection methods significantly influence the frequency of reported workplace violence against nurses.

Findings

Workplace violence was positively correlated with stress and inversely with work ability.

Direct questioning led to higher reported rates of violence compared to indirect methods.

Younger nurses reported significantly higher rates of physical aggression and all forms of violence.

Abstract

Background/Objectives: Workplace violence (WV) against healthcare workers (HCWs) is a major hazard all over the world. Prevention requires a reliable risk assessment. The rate of HCWs reporting a violent event varies considerably across multi-year retrospective studies compared to periodic surveys. We conducted a rapid observational study to demonstrate that data collection methods are more important than socio-cultural and healthcare organizational differences in determining the frequency of reported violence. Methods: In June 2025, in a cross-sectional observational comparison, we examined a total of 236 nurses divided into three groups: the first two were recruited online from Brescia (Italy) and Trollhättan (Sweden), while the third group was composed of Latium (Italy) nurses participating in a sleep health promotion program who answered the same questions on WV online. All the workers reported the frequency of violent incidents experienced in the previous 12 months using the Violent Incident Form (VIF), occupational stress using the Effort/Reward Imbalance questionnaire (ERI), and work ability via the Work Ability Score (WAS). Results: In the three samples, WV was correlated positively with stress and inversely with work ability (p < 0.01), while no significant difference was found between Italian and Swedish nurses in relation to the spot surveys. The nurses questioned directly about WV were significantly younger and reported significantly higher rates of physical aggression (28% vs. 5%, p < 0.001) and all forms of violence (73% vs. 20%, p < 0.001) than those questioned indirectly during the census of all the HCWs. In a multivariate linear regression model, the WV experienced and poor work ability were highly significant predictors of work-related stress (p < 0.001). Nurses who had experienced WV in the previous year had an increased odds ratio (OR = 8.94; Confidence Interval 95% = 4.43; 18.01) of reporting a state of distress. Conclusions: Experience has shown that specific questioning about violence—the commonest method used—encourages respondents to report violent events and may induce overreporting. This method also tends to involve younger workers who are more exposed to WV. On the other hand, prospective studies based on official reports may be influenced by underreporting. Monitoring WV during health promotion interventions included in occupational health surveillance could minimize both phenomena. Systematic studies and meta-analyses which rely mainly on “ad hoc” studies may be biased.

Figures